Open calls in doctoral studies

Supervisors: Kaire Torn, Kristjan Herkül, Martin Gullström

Coastal ecosystems play an important role in the global carbon cycle by absorbing atmospheric CO2 and thereby mitigating climate change. A large share of carbon is stored and sequestered in the sediments of underwater vegetated ecosystem. In the Baltic Sea, there have been relatively few studies on coastal carbon storage; previous studies have primarily been focused on seagrass meadows. However, some recent findings have indicated that other species of rooted vegetation and macroalgae also contribute significantly to carbon sequestration. Carbon storage exhibits considerable spatial variability, emphasizing the need for detailed site-specific studies to elucidate its spatial patterns along environmental gradients.

The primary objective of this doctoral project is to address critical knowledge gaps related to the patterns and processes of blue carbon in the coastal ecosystems of the northern Baltic Sea. The study aims to quantify the amount of carbon stored in coastal areas and to detect variations among species, communities, and along environmental gradients. Through the combination of point data on carbon content from different communities and distribution maps of macrophyte biomasses, a spatial map of carbon stocks across Estonian coastal waters will be generated. The results allow to define hot spot areas of blue carbon habitats and thereby benefit marine spatial planning processes. The results also enable the quantification of the total budget of coastal blue carbon including estimates of carbon storage in both living macrophytes and in seabed sediments. Based on primary production measurements, the species-specific carbon uptake capacity of common macrophytes species will be determined. 

Supervisors: Arvo Tullus, Priit Kupper

Norway spruce (Picea abies (L.) Karst.) is ecologically and economically one of the most important tree species in Central and Northern Europe. However, the adaptability of spruce to climate change is considered to be inferior compared to other main tree species in this region. According to climate change projections, the amount of precipitation in Northern Europe (incl. Estonia) will increase by 5-30% by the end of this century. An increase in extreme weather events (droughts, floods, heat waves) is also predicted. P. abies is a drought-sensitive tree species, and it is threatened by mass outbreaks of spruce bark beetle (Ips typographus) and fungal pathogens (e.g. Heterobasidion parviporum). Hence, the overall fitness of spruce is largely determined by its ability to cope with abiotic and biotic stresses whereas trees’ stress tolerance is generally believed to be higher in mixed than in monospecific stands. The primary
physical and chemical defence of conifers against biotic damagers is provided by resin in bark and sapwood. The chemical defence ability of spruce is also determined by the synthesis of specialised metabolites (incl. phenolic compounds) in the phloem of the bark and needles. As resin and polyphenolic compounds production requires energy investment, there could be a trade-off between tree growth and defence chemicals production. The existing knowledge of spruce’s defence capability under changing climate and different growth conditions is still incomplete, especially in northern part of its range. The aim of this PhD project is to analyse the defence capabilities of P. abies (resin flow, resin duct anatomical properties, content of phenolic compounds in bark and needles) at different forest site types and comparatively in monospecific and mixed stands. Climate change effects will be clarified based on dendrochronological analyses of core samples from 15 spruce forest plots and data from the FAHM climate manipulation experiment.

Supervisor: Riin Tamme, Meelis Pärtel

Autotrophic organisms (including vascular plants and algae) are fundamental for sustaining life on Earth. Adapted to diverse environments ranging from aquatic to terrestrial ecosystems, autotroph species exhibit varying habitat and environmental preferences. While broad evolutionary trends of autotrophic organisms’ habitats (e.g. water to land) are understood, the spatial distribution and phylogenetic patterns of habitat and environmental preferences remain unexplored. This PhD project aims to address this knowledge gap through three main research aims. Firstly, to review and consolidate data on autotrophic species’ habitat and environmental preferences, providing a global database. Secondly, it examines the global biogeography and phylogenetic diversity of vascular plants with different habitat preferences, environmental niches and specialization. Thirdly, the project extends to algae, analyzing their global distribution patterns and phylogenetic diversity across different habitat and environmental preferences. This PhD project contributes to a novel understanding of autotrophic organisms’ habitat preferences on a global scale. Understanding the biogeographic and phylogenetic patterns of autotrophic organisms’ habitat preferences is essential for predicting and mitigating the impacts of global change.

Supervisor: Kadri Koorem, Marina Semtšenko

In the current global change era, ecosystems face multiple stressors, such as the increasing frequency of extreme weather events and decreasing species richness, which influences the stability and functioning of ecosystems. At the same time, the human population on Earth is increasing rapidly. As a result, society faces an urgent challenge to support diverse and functioning ecosystems while maximizing food production. One of the options to tackle this problem is using crops that support biodiversity and ecosystem functioning. 

This PhD project will evaluate the capacity of different crop species to support biodiversity and ecosystem functioning. As a warming climate also offers the possibility to grow crop species that have so far been grown in warmer climates, these so-called novel crops will be the focus of this PhD project. Firstly, this project will evaluate the above-and belowground traits of novel crop species. Secondly, the growth of these crop species in combination with distinctive soil microbial communities in extreme weather events such as drought will be evaluated. Thirdly, the growth of selected novel crops will be tested in field conditions while also measuring parameters related to ecosystem functioning. 

In combination, the results of this PhD project will advance current knowledge about ecologically sustainable food production.

Supervisors: Sten Anslan, Leho Tedersoo

High-throughput sequencing tools are broadly used in biodiversity analyses, but the authenticity of the so-called rare biosphere remains an open question. The main analytical artefacts include sequence errors and chimeric molecules generated during polymerase extension in PCR, and signal conversion errors during the sequencing process. Although 50-90% of these errors can be eliminated during bioinformatics quality filtering, a significant proportion remains and continues to haunt biodiversity analyses by overestimating diversity or merging closely related species. In some datasets, most of the remaining artefacts can be easily noted by eye in multiple sequence alignments, but there are no automatized procedures to remove these. This situation fits well into the logic of Artificial Intelligence (AI) that can be supervised to automatically locate and “treat” such errors based on relatively simple models. This project is focused on utilizing AI for detection of Taq or custom polymerase errors, platform-specific sequencing errors and chimeric reads from quality-filtered metabarcoding datasets. It is expected to result in creating an analytical routine for high-precision quality-filtering analysis, which may potentially transform the way of such type of work.

Supervisors: Krista Takkis, Liis Kasari-Toussaint, Triin Reitalu

Supervisor: Tuul Sepp

Current urban ecology research has centred on birds and plants, but bats are considered excellent bioindicators of habitat quality. So far, studies of urban bats have mostly focused on community structure, but a focus on animal health allows for a better understanding of the features of the ecosystem that determine the suitability of habitats for wild species. This project is linked to a 5-year urban habitat restoration project (urbanLIFEcircles), where we plan to assess habitat quality not only by monitoring the diversity and abundance of wild species, but also apply a novel approach of condition-index based monitoring. We have chosen bats there as main indicator species, as they are present in both urban and natural habitats, are on the top of the food chain (bioaccumulators) and linked to both aquatic and terrestrial habitats. In this PhD project, we will apply different methodological approaches. First, we will conduct a citizen-science study to assess bat abundance and diversity in relation to urban habitat features. Bat vocalization data collected by volunteers will be analysed together with remote-sensing data to create a model predicting bat diversity hot-spots in urban areas. Secondly, we will collect samples from bats from habitats with differing levels of human disturbance, to assess their exposure to pollution, their parasite burden, and their health state. We will also assess the prevalence of Lyssa virus in Estonian bats, to fill a gap in the current knowledge. Finally, we will assess the feeding habits of migratory bats in different habitats based on a metabarcoding study. This project will give us a better understanding of health status of bats in urban habitats, and allow to give recommendations for designing more biodiverse and ecologically wholesome cities.

Supervisors: Asko Lõhmus, Ants Kaasik

In temperate zones, forests constitute the dominant potential vegetation, which is important as a reference for assessing human impact on terrestrial ecosystems and for strategic planning of ecosystem goods and services. While forest ecological studies have a long history, the prediction of future states of forests over regional scales has remained very general; this inhibits informed strategic forestry planning over mid-term time horizons. The UT Conservation Biology working group has recently developed a full-scale spatial simulation model “NextStand” for such future predictions (Kaasik et al. 2023 PLoS ONE). The current doctoral project will address the main uncertainties of forest futures, with broader objectives of (1) describing and explaining the predictability of future ecosystems in general; (2) improving the realism and diversity of “NextStand” model predictions; (3) predicting the dynamics of some actual ecosystem goods in the Baltic States (at least) in a year 2050 perspective. Specifically, the project will address: (a) through field studies – the possibilities to predict and model natural disturbance events, and the resulting forest structures and management responses, as stochastic processes in the spatial model; (b) model extension – modification of the NextStand model for Latvia, and incorporating realistic socioeconomic scenarios for the whole Baltic region (current parametrization of the model is based on Estonia); (c) compilation and testing of new output variables derived from the current stand description output, e.g. estimates of dead wood storage, habitat functions etc. The project provides a basis for new research hypotheses regarding the factors and consequences of future forest dynamics, and it opens multiple potential applications for spatial planning.

Supervisors: Jeffrey Malherbe Carbillet, Jaanis Lodjak, Tuul Sepp

Seabirds are one of the world's most endangered avian groups due to human-induced rapid environmental change, which is causing, among other things, a decline of suitable habitats. This decline increases the density of birds in the remaining suitable habitats, leading to a strong competition with conspecifics and representatives of other species that can reduce survival and reproductive success. Selection of a nesting site, which is crucial for improving the chances of a successful breeding event for a seabird, takes place at several spatial levels, starting from wider landscape scales and being finalised in a very small local scale of a few meters. Given that most seabirds are faithful to their breeding site, and the habitat loss is decreasing options for even the ones who would normally leave their colony of origin, small-scale nest location choice becomes increasingly more important. Within a breeding colony, habitat varies due to both landscape properties (micro-climate, vegetation, topography) and interactions with conspecifics and representatives of other species (social interactions, competition, predation), and this affects nest-site quality and reproductive output. While habitat modification and the use of low-quality nesting sites are known to reduce survival and reproduction in seabirds, the mechanisms through which these effects occur, and the potential for individuals to adapt through physiological and behavioural mechanisms remain unknown. Within this PhD project, we plan to use a longitudinal approach looking at nesting site selection in a colony of Common gulls (Larus canus) that has been monitored over the last 40 years, on an islet in the Matsalu National Park, in Western Estonia.  The density of breeding Common gulls, but also other species has increased over the past decades, making it a perfect colony to study the factors affecting nesting site selection and its consequences on fitness. The aim of the project are 1) To determine the factors affecting nesting site selection, such as previous experience, genetic components, and behavioural profiles 2) To identify the mediators between environmental conditions around a nesting site and reproductive success, considering glucocorticoid hormones as a candidate, 3) To evaluate the consequences of nesting site selection on survival and reproduction, also considering the potential to adapt among years through life-history strategies adjustments.

Supervisors: Kadri Runnel, Ivika Ostonen-Märtin

This ambitious PhD project (part of the larger FutureScapes initiative) aims to assess carbon accumulation potential in hemiboreal forests by concurrently analyzing decomposition in both above- (dead wood) and belowground (soil) carbon pools, identifying key biotic and abiotic drivers, and the potential climate change effects. Soil and dead trees represent over half of short to medium term forest carbon pools, and are vulnerable to forest management and climate change. The key innovation of this project lies in simultaneously assessing the drivers enabling carbon accumulation in these pools, which have in scientific literature so far been considered separately. This enables a holistic measurement of qualitative and quantitative changes in forest organic C accumulation pathways in relation to abiotic conditions and environmental changes (soil moisture and temperature) and planning climate-smart management which simultaneously mitigate carbon footprints while preserving biodiversity. Using a study setup established in 2023, the project will complement existing experiments by studying decomposition of additional substrates and analyzing data collectively. You will be able to do fieldwork in Estonian hemiboreal forests, learn modern lab techniques (qPCr and metabarcoding related), bioinformatics and statistics, all desired skills in the Academia. Outputs include a review on forest management strategies for carbon retention and case studies examining carbon accumulation in forests across abiotic gradients.

Supervisors: Sergei Põlme, Leho Tedersoo

Fungi are one of the key pathogens and root symbionts in forest and agricultural soils, and they are usually identified by sequence analysis of marker genes. In multiple practical occasions such as forest or cropland health surveys, it is extremely difficult and time consuming to link the hundreds of retrieved species-level taxa to their potential functions. Four years ago, we released the FungalTraits database that assigned fungal genera to specific trophic strategies (updated only for newly described genera). Unfortunately, the genus-level assignments are not good enough for the largest fungal genera that include individual species behaving as plant or animal pathogens, saprotrophs or endophytes. The main objective of this PhD project is to extend the FungalTraits database by including species-level functional information about trophic strategies for the large genera with variable lifestyles and to supplement this with information from genome sequencing projects and fungal community surveys equipped with rich metadata about substrate and putative host. The database has great importance for both education and research purposes in understanding the presence, diversity and distribution of fungi with different lifestyle.

Supervisors: Niloufar Hagh Doust, Kristel Panksep, Leho Tedersoo

Fungi and bacteria are the key symbionts in both terrestrial and aquatic ecosystems providing ecosystem services related to decomposition, nutrient acquisition, pathogenesis, etc. While both habitats separately have been well surveyed for biodiversity, little is known about the pattern of changes along the land to sea gradient and which environmental factors affect the coastal microbiome. In 2023, the TREC (Traversing European Coastlines) project was launched by a consortium led by EMBL to comprehensively survey the biological, geochemical and atmospheric properties of the land to sea gradient across Europe. This collaboration PhD project is set up to focus on the fungal biodiversity in leaf, soil, sediment and water samples across Europe. The main objective is to document the biodiversity of fungi in soil, shallow-water sediments and deep-water sediments across Europe and disentangle the underlying climatic, substrate chemical and anthropogenic parameters using state-of-the-art statistical analyses. The PhD project provides an important piece of a puzzle to holistic understanding of biodiversity and functioning of the coastal ecosystems. We also gain much better knowledge how the fungi in soils, sediments and water are taxonomically and functionally differentiated.

Supervisors: Vladimir Mikryukov, Sergei Põlme, Leho Tedersoo

Forest soils harbour diverse communities of fungi that are mainly driven by the dominant plant species and soil pH. Usually, closely related plant, animal and fungal species have a similar niche due to phylogenetic niche conservatism, but our analyses of Estonian forest soils indicate that congeneric fungal species may have strongly contrasting niches with respect to host plants, soil pH, and nitrogen concentration. Meanwhile, other species have very broad niches, which are difficult to explain. The project aims to investigate the patterns of edaphic and plant-related niche differentiation in saprotrophic and ectomycorrhizal fungi in Estonian forest soils, determine how different sequence clustering thresholds affect our understanding of fungal niche differentiation, assess the relative importance of evolutionary conservatism and allopatric vs. sympatric speciation, and identify the typical genomic mechanisms underlying niche differentiation in closely related fungal species. The project includes compilation of published sequence data from Estonian forest soils, sequence clustering at different sequence similarity thresholds, constructing species distribution models and phylogenetic trees for ancestral state reconstruction, as well as using comparative genomics to explore genomic differences between species pairs with shared and divergent niches. 

Supervisor: Leho Tedersoo

Microorganisms such as fungi and bacteria form diverse communities in grassland ecosystems, with apparent specialisation by vegetation types, soil nutrients and pH. However, at the ecosystem level, there is no holistic understanding to what extent microbial communities overlap within grassland habitats and compared with forest and cropland ecosystems. This project aims to determine habitat specificity in different fungal and bacterial functional groups and the effect of sampling strategies on the efficiency of recovering microbial diversity. The project uses cutting-edge molecular methods for identification, bioinformatics and statistics, including taxon-specific primers. For both fungi and bacteria, marker genes of different length are used for comparing their efficiency in future studies. The potential importance is very high from the microbial ecology perspective, because knowledge about the relative habitat preferences of fungi and bacteria enable to understand how broad are the microbial niches and how this can be used in the conservation perspective. 

Supervisors: Urmas Saarma, Ants Tull

Pathogenic parasites exist in all ecosystems and majority of them are zoonotic, i.e. transmitted between animals and humans. Due to a limited knowledge of the zoonotic parasite communities in pets, infections can often go undiagnosed and untreated, both in pets and their owners. Therefore, it is vital to have detailed knowledge of parasite communities. Since the complex interaction between the mammalian host, parasites, and microbes is an important factor of mutual fitness, it is also vital to investigate to what extent the host’s parasites and microbiome are in the mutual interplay. Cats are known to transmit many pathogenic parasites and microorganisms to humans, some can cause severe diseases or be even life-threatening, especially to risk groups like children, pregnant women and immunocompromised individuals. The aim of this project is to analyse parasite communities of cats, especially the pathogenic taxa, their variability, and infection intensity. We also aim to evaluate relationships between the occurrence and infection intensity of different parasites, as well as the relationship of parasites and the microbiome. For this, we will collect fresh faecal samples of cats from 800 individuals originating from both rural and urban environments. The parasite and microorganism communities will be analysed by using DNA metabarcoding. Different classes of parasitic helminths (cestodes, nematodes, trematodes) and microorganisms (bacteria, fungi, protozoa) will be identified. 

The innovative aspects of this project are: 1) analysing the parasite communities and the intestinal microbiome of cats; 2) evaluating the relationships between the occurrence and intensity of different parasites accumulated from various environments; 3) investigating mutual interactions of parasites and microorganisms. 

This project is the first one to provide a holistic overview of the parasite and microbial communities of cats and the dynamics between them. Since zoonotic parasites and pathogenic microorganisms can seriously affect human health, the new knowledge derived from this project will be essential to raise public awareness of pathogens that cats can transmit to their owners and caretakers, as well as to design strategies to minimise the risk of infection to humans and to improve prevention and diagnosis rates. 

The doctoral student is admitted to a student's place of study, full-time, the nominal study period is four years. An employment contract for a junior researcher is not concluded. The planned start of study and research is September 2, 2024.

Supervisors: Maris Hindrikson, Egle Tammeleht

The study of wolf-dog hybrids is of the utmost importance as the increasing presence of such hybrids could have a major negative impact on Europe's fragmented wolf populations. To date, wolf-dog hybrids have been identified in all European countries where wolves are present, but there is very little published research on wolf-dog hybrids. In order to better understand the impact of hybridisation on current European wolf populations, the PhD study will determine the level of hybridization and introgression in European wolf populations. 

The work will also compare the genomes of dog-wolf and dog-wolf hybrids to understand patterns of hybridisation at the chromosomal level. In addition, analysis of mitogenomes and partial Y-chromosomes will also be carried out to investigate the differential contribution of maternal and paternal lines to the genomes.

The resulting thesis will provide an overview of the introgression of European wolf populations, wolf-dog hybrids and patterns of hybridization, and the impact that continued hybridization may have on the wolf-dog populations in Europe.

Supervisors: Rasmus Palm, Enn Lust, Thomas Thomberg

Synthesis of chemically pure highly microporous carbon materials is of high interest for a multitude of applications, one of which is the adsorption and separation from mixtures of energetically valuable gases. Electrospinning will be used to prepare nanocluster-doped polymer membranes, where a nanocluster-forming agent has been introduced during the electrospinning process. These doped polymer membranes will be used to prepare highly microporous carbon fibres through carbonisation. The project focuses on the use of different electrospinning process parameters, activating agents, and carbonisation routes to obtain chemically pure carbon fibres with optimal porous structures for hydrogen storage and separation applications. The prepared polymer membranes and carbon fibres porosity, structure, composition, fibre morphology, and fibre size distribution will be characterised. The carbon fibres' hydrogen storage and separation capability will be determined through adsorption measurements with different gasses. Neutron scattering methods will be used to investigate the hydrogen confinement and mass transfer properties of the microporous carbon fibres and to characterise the porous structures. New optimised synthesis routes for functional microporous carbon materials will be developed and the synthesised materials will be used as model and technical materials for explosion-risk free gas storage and separation applications.

Supervisors: Ivo Leito, Valter Kiisk

pH is the most widely used acidity measure but currently, rigorous, reliable and widely accessible pH measurement is only possible in bulk solutions and is impossible at phase interfaces/surfaces. This project aims to create an experimental approach to enable rigorous pH measurement at phase interfaces comprising aqueous (or aqueous-organic) and non-aqueous media (non-polar solvent) or gas (air), relying on the unified acidity scale (pHabs) and to apply it to a set of scientific problems where pH at interface is important. The measurement approach for pH measurement at interfaces will be based on the measurement of pH differences between interface and bulk solution by a set of pairs of ratiometric fluorescent pH-sensitive molecular probes.

Supervisors: Prof. Ago Rinken, Dr. Edijs Vavers

The Sigma-1 receptor (Sig1R) is an intracellular chaperone protein that also plays a role in modulating the release of extracellular vesicles (EVs). We have implemented the MultiBacMam expression system to produce multicolour vesicles and developed a sensitive Total Internal Reflection Fluorescence (TIRF) microscopy-based assay system to detect and monitor EVs. This enables us to further understand the properties and functions of Sig1R-containing EVs. The main objectives of this project include: Characterising EVs that contain Sig1R; Determining the factors that influence the formation and properties of these EVs; Developing a tool for Sig1R ligand screening using fluorescence-based techniques. Through this project, we aim to gain insights into the role of an intracellular chaperone in intercellular communication, provide innovative options for Sig1R ligand screening, and enhance our understanding of Sig1R-related mechanisms in cellular biology.

Supervisors: Dr. Anni Allikalt, Dr. Tõnis Laasfeld, Dr. Dmytro Fishman

G protein-coupled receptors play a crucial role in cell communication and neurotransmission. Dysfunctions in signal transduction are associated with various diseases, making these proteins important drug targets. The broader aim of this research is to investigate the genetic variability of G protein-coupled receptors in the human population. Existing genomic data will be analyzed, and genetic mutations will be selected for further experimental testing. In these experiments, a comparison of the interaction of already approved drugs with wild-type or mutated receptors will be performed. The main objective of this doctoral project is to analyze and manage a large amount of data collected from various types of assays so that the results are accessible and understandable to other researchers. The obtained data also allows the development of QSAR (Quantitative structure-activity relationship) models to predict such experimental values virtually in the future. Since one of the main test methods of the project is fluorescence microscopy, the focus of the doctoral project is on optimizing the analysis of microscopy images. This would make the analysis method more sensitive and enable the detection of small differences between different receptor variants.

Supervisors: Koit Herodes, Ivari Kaljurand

There are several thousands non-natural fluorinated compounds (FOCs) that have been amply demonstrated to be present in many parts of the environment. The majority of them are either potentially or demonstrably toxic and persistent. Thus, it is important to be able to reliably determine their presence and concentration. Current methods have significant limitations in terms of selectivity, identification and quantification. The aim of the work is to develop a synergistic platform involving LC-MS and 19F NMR, whereby both techniques exploit their strengths and mutually compensate for their weaknesses. LC-MS will provide sensitive detection and identification, while state-of-the-art 19F NMR – cryoprobe, hyperpolarization (SABRE, Photo-CIDNP), etc – will provide the comprehensive picture of FOCs together with accurate quantification and contribute to identification. This will result in obtaining the most comprehensive picture (and database) of FOCs in aqueous samples that is available to date.

Supervisors: Jaanus Harro, Mati Karelson, Margus Kanarik

Recently it has been shown that mRNA methylation of the amino group at the N6-position of adenosine (m6A) affects the translational efficacy and may be involved in the pathogenesis of stress-related disorders. Thus, the enzymes responsible for mRNA methylation regulation also offer a novel target for pharmacological intervention – prof. M. Karelson and colleagues have developed several small-molecule drugs that effectively modulate the activity of methyltransferases and demethylases. The first-in-class METTL3/METTL14 methyltransferase activator CHMA1004 efficiently protects dopamine neurons from neurotoxicity in rats, as does the inhibitor of the demethylase FTO. Systemic acute and sub-chronic administration of CHMA1004 has activating/anxiolytic/antidepressive effects on rat behaviour. This doctoral project seeks to: clarify the pharmacokinetic properties and the sex- and dosedependent effects of several mRNA modifying drugs on the behaviour of rats in tests of locomotion, anxiety- and depression-like behaviour; determine the brain regions where the epitranscriptomic modifiers and neurotrophic factors cause a change in mRNA methylation levels (as analysed by UHPLC with MS detection); determine the genes and biological pathways where transcripts are affected by the administration of mRNA modifiers and neurotrophic factors (as measured by RT-PCR and RNA m6A sequencing); and determine the effect the mRNA modifying drugs on monoamine neurochemistry (as measured by HPLC with electrochemical detection). We expect increased m6A methylation in anxiety-related brain regions to be anxiolytic and aim to identify novel molecular targets for drug development.

Supervisors: Siim Salmar, Jaak Järv, Siobhan O. Matthews

Mechanical processing of pine wood in Estonia is world class, an example of which is Barrus AS, a leading European manufacturer of pine finger-jointing and glued laminated timber components in South-East Estonia, targeting door and window manufacturers. However, the production process leaves over 200 000 m3 of residues per year (over 10 million m3 in Estonia and the Baltics combined), which are mainly used in wood pellet production for cheap energy. Pine wood is rich in extractive compounds such as lipids, resin acids, lignols, and more, which makes pine biomass a non-favourable feedstock for use in biorefining technologies. The Laboratory of Wood Chemistry and Bioprocessing at the University of Tartu has planned a Ph.D. project in cooperation with Barrus AS, which focuses on the development of novel refining technologies for pine wood with the aim of finding the most optimal technological and economic solutions for wood biorefining. The use of supercritical fluids for the extraction of bioactive substances has proven itself in the pharmaceutical and food industries. In this project, we are collaborating with one of the world's leading developers of supercritical fluid technologies, SCF Processing Ltd., to find ways of pre-treating pine wood for the recovery of extractive compounds. The project will test different ways of using supercritical fluids, resulting in the necessary findings to upscale to a pilot reactor design at the University of Tartu. Valuable applications for pine wood extractive components and lignocellulosic biomass will be analysed and found to provide input for integrated pine wood biorefining.

Supervisors: Rasmus Palm, Enn Lust, Thomas Thomberg, Angélica María Baena Moncada

Metal-organic frameworks (MOFs) include a wide variety of meso- and microporous structures with the potential to incorporate different metals and heteroatoms. Carbonisation of MOFs yields highly porous carbon materials with uniformly distributed catalytically active functional centres. The aim of this PhD project is to synthesize and carbonise MOFs yielding highly functional catalytically active model carbon catalyst materials. The MOF derived materials will be used to investigate the effect of different active sites in a carbon material to the adsorbate/adsorbent interactions important for energy storage and energy conversion applications. The focus of this PhD project will be to synthesise, carbonise, and perform physical characterisation of MOF-based materials. This will include the optimisation of the MOF synthesis and carbonisation routines to obtain highly porous materials with well-tuned highly accessible active sites. The fine-tuned active sites would be of interest for the investigation of hydrogen adsorption and hydrogen spill over effect, as a catalytical support for nanoconfined hydrides, and as a catalytic site for oxygen reduction reaction. The PhD project will involve the physical characterisation of the prepared materials with various methods like gas adsorption, X-ray diffraction, and Raman spectroscopy. In addition, neutron scattering experiments will be performed to investigate the hierarchical porous structure and processes of interest for energy storage and conversion applications on these materials.

Supervisor:Arun Kumar Singh

Hand-crafting rules of robot planning and control through the first principles of optimality, controllability often leads to conservative performance in dynamic and unstructured environments. On the other hand, data-driven, end-to-end approaches that directly map sensory inputs to control actions have failed to scale beyond control laboratory scale experiments owing to the concerns of repeatability, robustness and safety. The overall objective of the PhD project is to solve the fundamental challenges associated with data-driven planning and control; (i) scarcity of real-world training data and (ii) lack of performance, safety guarantees and explainability in the learned model. The PhD project will focus on navigation of single and multiple agents in heterogeneous environments. Specifically, this includes navigation on different terrain conditions and in environments populated with agents with diverse/heterogeneous capabilities (e.g crowd, urban driving). To this end, the project will develop physics-aware priors for modelling wheel-terrain and agent-agent interaction based on differentiable optimization and ODEs.

Supervisor: Mozhgan Pourmoradnasseri

Understanding and modelling human mobility behaviors is essential for urban planning and transportation management. This research addresses the challenge of generalizing mobility models from one city to another, with a specific focus on active mobility and micromobility. Current deep learning models, when trained on a specific city, struggle to adapt effectively to different urban environments. The thesis explores the limitations of existing models and proposes innovative approaches inspired by advances in deep learning methods and aims to enhance model generalization across diverse geospatial regions. With a focus on active and micromobility, the ultimate goal is to develop city-agnostic mobility models that can transcend geographical boundaries, fostering advancements in real-world applications and urban mobility planning.

Supervisor: Kaur Alasoo

A fundamental challenge in human genetics is deciphering the molecular consequences of genetic variants associated with human complex traits. A small fraction (<10%) of these variants alter protein function by directly modifying the coding sequence of a protein. More commonly, these variants regulate how the protein-coding template is assembled via RNA splicing (‘splicing variants’) or how much protein is expressed (‘expression variants’) in a difficult to predict manner. A gold standard for identifying the effects of splicing and expression variants have been quantitative trait locus (QTL) studies that directly measure the effects of these variants in a natural population. We have used this strategy to build the widely used eQTL Catalogue database - the largest compendium of uniformly processed human gene expression and splicing QTLs. However, QTL studies are necessarily limited to characterising only a very small fraction of all possible genetic variants that can be observed in a small population. To scale variant interpretation to hundreds of thousands of associations discovered in existing and upcoming million-scale biobanks, there is a significant need for well-calibrated, interpretable machine learning models that can accurately predict the consequences of any genetic variant. This project will combine state-of-the-art machine learning models with comprehensive gold standard datasets to make predictive regulatory genomics a reality.

Supervisors: Mark Fišel, Elizaveta Yankovskaya

Machine translation is undergoing rapid transformation due to the discovery of emergent abilities of large language models. As a result, the previous state-of-the-art approach of isolated sentence-by-sentence translation (still running behind Google Translate, Neurotõlge and other similar services) does not have the flexibility required to deliver more reliable and adaptive automatic translation. This project will explore the development and exploitation of emergent abilities in translation models with the goal of enabling rapid adaptation to user feedback, terminology integration and document-level context consideration. In this project we will focus on high-to-mid-resourced languages (Estonian, English, German, Russian, Ukrainian) as well as low-resourced languages from the Finno-Ugric language family.

A prospective candidate should have a strong background in NLP and math with experience of text processing as well as training and evaluating neural networks.

Supervisors: Piret Luik and Marina Lepp

The research aims to address the significant gender disparity in participation in the Informatics Olympiads, where the percentage of girls in the final round of the Estonian Informatics Olympiad is about 10% from 2015 to 2023. The research questions aim to understand the reasons behind this low participation and explore ways to increase the proportion of female participants.

To answer research questions, a mixed methods study will be conducted. At first, the study reviews relevant literature and curricula of the different countries organizing the Informatics Olympiads, and collects data from different countries regarding support measures. Secondly a questionnaire among university informatics students in Estonia is used to explore factors influencing students' participation in informatics contests. Qualitative data will also be gathered through interviews with girls who have participated in Informatics Olympiads.

By exploring factors influencing female participation in informatics contests and understanding the impact of the experience on them, the research aims to contribute to encouraging and supporting women in IT. The study emphasizes the importance of gender equality in driving progress and innovation in the field of informatics. The findings could have implications for designing effective support measures and strategies to increase the representation of girls in the Informatics Olympiads.

Supervisor: Mohamad Gharib

In our rapidly evolving technological landscape, novel technologies are constantly integrating into our daily lives. Despite their aim to simplify our lives, these technologies also expose us to an array of security and privacy threats. To address this concern, a range of security and privacy solutions, encompassing various techniques and mechanisms, have been developed to safeguard users from diverse threats and attacks.   However, a significant body of research has highlighted a prevalent issue: many of these solutions often fall short of their intended goals due to a lack of user understanding of their proper usage. Consequently, the effectiveness of security and privacy solutions relies heavily on end users being able to utilize them correctly without impeding their primary tasks.  For more than two decades, researchers in the field of 'Usable Security and Privacy (USP)' have been dedicated to resolving this issue by empowering users to make well-informed decisions regarding security and privacy. However, existing solutions continue to struggle to achieve this objective.   Many of these solutions are designed with the average user in mind, often struggling to effectively balance the interplay of security, privacy, and usability. Challenges arise from inadequate capture and analysis of USP requirements, leading to these requirements often being overlooked in the final solution, resulting in systems that are barely usable. A potential solution to this issue lies in aligning the design of USP solutions with the mental models (MMs) of their intended users. In essence, this means designing solutions that cater to the user's ability to make well-informed and appropriate decisions.  However, currently, there is no agreed-upon approach for incorporating MMs into USP, nor is there clarity on how to accurately capture, represent, or rectify these models to aid users in their decision-making process. This thesis aims to address the aforementioned challenges by proposing a comprehensive framework for designing USP solutions that align precisely with diverse users’ MMs, irrespective of their experience, demographics, etc. The framework encompasses four key components: (1) a modeling language providing concepts and constructs to capture both USP requirements and users’ MMs; (2) a set of analysis techniques to ensure the correctness and consistency of the models; (3) an approach to derive USP design solutions from the models and evaluate these solutions with end-users; and (4) an automated tool-support to assist designers during the USP solution design process. The efficacy of the framework will be validated through its application to two distinct case studies spanning different domains.

Supervisor: Helger Lipmaa

Zk-SNARK is a cryptographic construction that allows the prover to prove that some computation has been done correctly without revealing private data; that is, to perform verifiable computation. A key goal, allowing scalability, is for the verification to be faster than the original computation; this makes zk-SNARKs appealing even in applications that do not require privacy. Various companies have implemented and used zk-SNARKs to perform verifiable computations at scale. Unfortunately, the main bottleneck is the prover's efficiency. This project aims to improve the state of the art by constructing more prover-efficient zk-SNARKs that allow more efficient verifiable computation. We study and perform cryptographic research on various aspects of this problem: for example, code and lattice-based zk-SNARKs and efficient incrementally verifiable computation. The student will collaborate with the supervisor's existing network in Estonia and abroad.

Supervisors: Ulrich Norbisrath,Steffen Manfred Noe

This doctoral thesis proposes extending the IoTempower framework, traditionally used in home automation and agriculture, to forestry management. The project will enhance forest monitoring and environmental accounting by leveraging the SMEAR research station network. A significant innovation is the integration of Artificial Intelligence (AI) for anomaly detection and analysis of visual data from satellites and drones, improving forest ecosystem monitoring.

The research focuses on making advanced IoT technology accessible using affordable sensors and microcontrollers. It also plans to implement forestry research inside the IoTempower teaching method at the University of Tartu and the University of Life Sciences to develop IoT and environmental monitoring expertise.

Combining IoT, AI, and environmental science, the project introduces a new approach to environmental accounting in forestry. Modernizing the SMEAR network with IoTempower and using 3D printing for sensor mounts and drone adaptations are key aspects of this interdisciplinary approach.

This thesis aims to advance forestry management by integrating innovative technologies, contributing significantly to IoT and environmental sciences. It underscores a commitment to sustainable environmental practices and expanding the impact of IoT technology in society.

Supervisors: Kerli Mooses, Raivo Kolde

Adherence to prescribed treatments determines largely the success of managing a chronic disease and avoiding adverse outcomes. The reasons for poor adherence are multifaceted ranging from personal to healthcare system level factors. To tackle the poor adherence rates we must better understand the determinants of the adherence. In current doctoral project we will utilise a real-world and representative data from e-prescriptions, electronic health records and insurance claims which allows a valid and cost-effective estimation of drug adherence and its determinants at scale. The doctoral project which will focus on:

1. Enhancing the data quality of OMOP database containing information about prescriptions through creating a tool which helps to evaluate and improve the prescriptions data quality of OMOP database.
2. Identifying person-level determinants affecting the drug adherence and their effect on health outcomes. For this we will include different health variables and apply advanced analysis methods to identify the person-level effects which will be used in later analysis to predict non-adherence, evaluate survival and hospitalisation.
3. Expanding research results into an international context. The established research methodology for systematically describing factors influencing treatment adherence is applied for conducting international studies. The goal is to create a comprehensive knowledge base regarding patient treatment adherence and the factors influencing it across different countries.
4. Dissemination of the results which will be mostly carry out through international conferences and research papers.

Supervisor: Somnath Banerjee

Large language models (LLMs) have proven to be highly effective at various natural language processing tasks, including answering complex questions, machine translation, etc. The text-generating powers of LLMs have advanced to a point where they are on par with human writers, thanks to their rapid development. Due to LLMs' robust generation skills, people now find it difficult to distinguish between textual content created by LLMs and humans, which has led to the development of complex issues such as LLM-generated misinformation. Misinformation (such as fake news and rumours) has been a longstanding and serious concern in the contemporary digital age, particularly in uncompromising industries like healthcare and finance. In the battle against misinformation, LLMs have generally been a double-edged sword, posing both new opportunities and challenges. On the one hand, LLMs' extensive world knowledge and robust reasoning abilities indicate their ability to completely transform the conventional paradigms for attribution, detection, and intervention with misinformation. On the other hand, the ability of LLMs to produce human-like content can be easily used to create misinformation, whether intentionally or accidentally. In these circumstances, this thesis addresses the detection and attribution of LLM-generated misinformation. This research proposal aims to protect human society against the negative impact of LLM-generated misinformation from a cognitive perspective. The importance of this proposal is two-fold: social as well as technical, that is investigating an important societal issue and contributing from a technical perspective by exploiting the state-of-the-art approaches (including LLMs ) to mitigate LLM-generated misinformation. In addition, this study investigates suitable approaches in low-resourced settings within the scope of Fino-Ugric languages (e.g., Estonian) and LLM-generated misinformation.

Supervisors: Anastasija Nikiforova, Kuldar Taveter

In view of the unprecedent value of data as an input for any data-driven activity or process, in times of democratization and globalization of AI, competitiveness largely depends on the availability of data. Over the past two decades, there has been increased interest in turning data into a public good, predominantly G2C and G2B data governance models, with a growing awareness of the need to combine it with other models such as B2G, B2C, and C2B.

Despite some progress being made in past years, today the future of public (open) data movement remains uncertain, where there is now a certain stagnation that the community is calling for an end to. Thus, the topics of public data ecosystems and data sandboxes are becoming popular.

While at the level of national open data ecosystems there is at least some degree of agreement on some of the characteristics that they should meet, this understanding is limited to the isolated concept of a national OGD portal and, sometimes, the stakeholders involved in communicating with it, where the public data ecosystem must consider the local level of cities. Understanding of the entire data ecosystem, which includes the (smart) city data ecosystem, is limited, with most designing and maintaining it using an ad-hoc approach that does not necessarily meet the needs and expectations of the quadruple helixes, where government interests dominate, often ignoring other stakeholders. This is especially true in relation to the trend of city smartification, which is associated with an exponential increase in data collection, where cities are not harnessing the potential of data generated by citizens, industry, academia, and government.

The aim of the project is to develop an inclusive and sustainable public data ecosystem. Specifically, the project will investigate existing public data ecosystems, their components and the relationships between them, to identify opportunities to (re)design them using data, process, user and technological transformations, and the magnitude and impact of invented (or projected) changes. As a result of this exploration, a public data ecosystem will be built around the corresponding platform(s) serving as a nexus between the municipal administration and its residents, fostering transparency, engagement, and innovation. The project will explore the hypothesis that a combination of ecosystem, platform (involving UX expert review and competitive analysis), and orchestration approaches coupled with 4I (the synergy of Artificial Intelligence with Data Intelligence, Collaborative intelligence, and Embodied Intelligence) contributes to a better understanding of the current data ecosystem, identifying its weaknesses and further eliminating them through the application of appropriate heuristics. 

Supervisor: Arnis Paršovs

Smart home systems, wearable devices, security systems, smart health devices, toys, and various other radio-connected devices are examples of equipment susceptible to hacking and privacy issues. These devices have the ability to monitor and collect sensitive user data over time, transmitting it through potentially insecure short-range communication technologies. Without proper encryption and strong authentication mechanisms, the data can be intercepted when the radio equipment transmits or receives data.

The goal of this thesis is to develop and improve methodologies for analyzing the security of radio-connected embedded devices. Specifically, the research will focus on popular embedded security devices available in the market. To achieve this goal, the doctoral project will integrate methods from electronics, radio signal analysis, side-channel attacks, cryptanalysis, and reverse engineering of microcontroller firmware. 

By combining these interdisciplinary approaches, we aim to enhance the understanding of potential security vulnerabilities in these devices and devise effective measures to mitigate the risks associated with their usage. Ultimately, this research seeks to contribute to the advancement of secure radio-connected embedded devices, safeguarding user data and promoting a safer and more reliable technology ecosystem.

Supervisor: Meelis Kull

In a world increasingly reliant on artificial intelligence (AI), ensuring that machine learning models accurately assess their confidence levels is crucial. This is especially true in areas such as healthcare and autonomous driving, where inaccurate risk assessment can lead to harmful decisions. Current strategies to mitigate overconfidence, such as Bayesian methods and post-hoc calibration, frequently struggle with complex tasks and dynamic environments due to their inherent limitations in reasoning and adaptability. This doctoral project seeks to tackle overconfidence by integrating probabilistic machine learning with generative AI. Generative AI systems (such as large language models) provide capability for identifying risk factors, modelling their dependencies, and estimating the probabilities of these individual factors. These capabilities can be complemented by probabilistic machine learning methods, including inference with graphical models, Bayesian neural networks, and post-hoc calibration. The planned methods have wide applications, including text classification, reasoning tasks, and real-time risk analysis, for example in autonomous driving. As a result of this work, it will be possible to build more reliable and trustworthy AI systems.

Supervisors: Faiz Ali Shah, Dietmar Pfahl, and Kallol Roy

Vulnerabilities in software systems can lead to security failures when exploited, compromising confidentiality, integrity, or availability. These vulnerabilities often stem from low-quality source code. Traditional static analysis tools for security code review, although scalable and useful early in the development lifecycle, have limitations. They might miss some vulnerabilities (false negatives) or incorrectly flag non-vulnerabilities (false positives), which can waste time and resources. To enhance effectiveness, a static analysis tool should accurately identify all vulnerabilities with minimal false positives and provide guidance for fixing them.

Recently, Large Language Models (LLMs) like ChatGPT and CodeLlama2 have shown the potential to detect vulnerabilities based on coding style attributes. This project proposes a novel approach using Code Language Models (CLMs) for detecting and mitigating code vulnerabilities. These open-source LLMs are fine-tuned on source and binary code and natural language artifacts, transforming into CLMs. These CLMs offer executable test cases for validating identified vulnerabilities. Employing transformer neural network models, CLMs can detect vulnerabilities by analyzing code similarity semantically, thanks to training on language corpus, source code fine-tuning, and innovative embedding methods. This approach allows CLMs to achieve semantic understanding, enhancing their ability to detect and explain code vulnerabilities beyond mere syntax.

Supervisors: Jaak Vilo, Sulev Reisberg

This Ph.D. project addresses the challenges posed by the complexity and temporal inaccuracies inherent in real-world health data, aiming to enhance the utility of real-world evidence as a cost-effective alternative to traditional clinical trials.The primary focus is on identifying generalized health events, determining their chronological order, and establishing similarities among temporal event sequences. Three articles outline the methodology: the first employs context-based analysis methods to group raw events into higher-level generalizations, the second strives to establish the true order of health events, and the third expands on the second article's methods to identify similarities among event sequences, as well as provide methodologies for alignment of patient trajectories. Through the development of analysis methods and new software packages, the student demonstrates the applicability of these approaches on Estonian data and any other health dataset transformed to OMOP common data model, contributing significantly to the advancement of real-world evidence utilization in medical research.

Supervisor: Dietmar Pfahl

Automated Driving Systems (ADS) are difficult to test, if they are at the levels 4 or 5 of automation (i.e., the ADS must function safely within a defined Operational Design Domain (ODD) and detect when it is leaving the ODD – completely autonomously, without relying on a human driver sitting in the car or remotely). Due to the nature of the software used in ADS, i.e., artificial intelligence in charge of perception, localization, route planning, and control of the physical car, testing of the ADS would require the observation of the ADS behaviour in an almost infinitely large array of scenarios to ascertain a required safety level. 

The focus of this thesis project is on defining a method that helps safety-engineers develop, define and implement scenarios in a simulator such that a) meaningful comparisons with safety-levels of human drivers in defined situations and under defined conditions can be made to generate comparative safety profiles, and b) synergies with on road testing can be achieved (to minimize cost of testing and allow for testing situations that are difficult to test on the road).

Supervisors: Jaak Vilo, Sulev Reisberg

Real world health data poses major opportunities for analysis of various aspects of health and disease in people. IN this thesis the ideal candidate possesses or develops algorithmic, statistical and visualisation competences that are needed for analysis of such data. Firstly, rapid characterisation of defined or queried patient cohorts is needed that highlights the main differences to the general population or compares directly to control groups. Secondly, we will develop methodologies for further sub-clustering the cohorts in respect to characteristics of their health statuses. One of the added complexities stems from the temporal aspects of event trajectories, as well as the continuous nature of the underlying measurements in the data. Lastly, the aim is to develop end user tools for querying and visualisation of such data. 

Supervisors: Marlon Dumas, Fredrik Milani, David Chapela de la Campa

Existing approaches to business process optimization are largely focused on the strategic and tactical levels of a process. Their purpose is to help business analysts to identify changes to a business process that will have an impact in the medium term, with a timeframe in the order of weeks or months. However, optimizations at the operational level, i.e., runtime interventions that affect the ongoing executions of the process, can also have a high positive impact on the performance of a process.

Existing operational process optimization techniques focus on solving resource allocation and scheduling problems, working under the assumption that the process does not change or changes little over time. In practice, however, business processes are subject to frequent changes. Therefore, a more "adaptive" approach to operational optimization is required.

This doctoral project will develop techniques for adaptive business process optimization at the operational level. To achieve this goal, the first objective is to develop methods to rapidly detect and diagnose performance changes in a process. The second objective is to develop methods to prescribe interventions to optimize the performance of the process in light of the detected changes. Finally, the third objective is to reliably quantify the impact of the proposed interventions by means of short-term simulation.

Supervisors:Marlon Dumas, Fredrik Milani, David Chapela de la Campa

Existing business process optimization approaches require high levels of manual effort from experienced business process analysts to explore the available business process data, and to design and evaluate improvement options. As a result, current business process optimization approaches are not scalable. 
This doctoral project will address this gap by combining process optimization techniques with generative AI methods, specifically Large Language Models (LLMs). The expected outcome is an LLM-based method and tool that takes as input process optimization objectives and questions posed by business analysts and engages in a conversation with these analysts to help them understand how to achieve their objectives. The tool will rely on business process execution data, business rules, textual documentation of organizational policies, and other inputs, to identify and formulate changes that can be made to a business process to achieve the optimization objectives posed by a user. The tool will also rely on business process simulation and optimization techniques to automatically explore spaces of optimization options based on the user inputs. 
The resulting method and tool should reduce the time required to define optimization strategies for a business process, while requiring less expertise than existing methods. This hypothesis will be tested via case studies and user studies. 

Supervisor:Raul Vicente Zafra, Jaan Aru

This research proposes a novel approach to understanding the formation of hexagonal grid patterns in grid cells, a key component of spatial navigation in mammals. By simplifying neural field models into wave equations, we aim to bridge the gap between neural dynamics and mathematical pattern formation theory. This methodology promises to reveal the underlying mechanisms of grid cell patterning, offering insights into the complex interplay between neural activity and spatial cognition. Through a combination of theoretical development, computational simulations, and empirical data analysis, this project seeks to provide a comprehensive model of grid cell function, with potential applications to general cortical theory and extending to artificial intelligence of spatial representations.

Supervisors: Marek Oja, Kerli Mooses

Clinical guidelines (CG) and care pathways (CP) aim to enhance the quality of care, increasing patient safety and ensure the effective use of healthcare resources. The real-world data from healthcare provision claims, electronic health records and e-prescriptions can provide more thorough understanding about the actual situation as well as trends over time providing an invaluable input both to the development and evaluation of implementation of the CG and CP. This in turn will help to improve the quality of healthcare services. 

The current doctoral project is aimed at improving the secondary use of Estonian health data in the development, monitoring and evaluation of CG and CP. The focus of the PhD project will be on the development of standardised reusable solutions (e.g. R package) to easily extract and visualise the information for CG and CP development and evaluation together with identifying deviations from CG or CP costs to the healthcare system. This project will utilise a database containing all health events from EHR, healthcare provision claims and e-prescriptions for a representative sample of Estonian population and which has been transferred to OMOP common data model (CDM). This in turn ensures the applicability of created solution in international scale.

Supervisor: Raimundas Matulevičius

AI/ML is applied across domains (self-driving cars, finance, healthcare, etc.) where various stakeholders and systems communicate sensitive data and decisions. Thus, protecting the data and information used in the AI/ML systems is instead a necessity rather than an option. This thesis will explore how AI/ML methods can be used for defensive and offensive security. It will also investigate the security risks to the AI/ML systems and how to mitigate them. Finally, it will consider what security requirements should be set to secure AI and ML systems and how security countermeasures can help manage the risks associated with AI/ML technology.

Supervisor: Dmytro Fishman

This project's overarching goal is to effectively utilize available biomedical imaging data by leveraging the strengths of existing deep learning approaches and developing new ones through innovative strategies. It aims to address the limitations of traditional CNN models, like U-Net, in processing complex cellular images and the high computational demands of transformer models. By integrating the advantages of CNNs and transformers, we propose a hybrid model approach to enhance both semantic and instance segmentation capabilities. Early work by Illia Tsiporenko on the Swin-UperNet model, incorporating deconvolution layers, demonstrates improved accuracy while maintaining efficiency. Additionally, we aim to expand our methodology to include a multimodal approach with CLIP, integrating textual meta-information with image data for a more comprehensive analysis. This initiative aims to transform biomedical imaging analysis by creating efficient, adaptable models that make the most of limited data resources.

Supervisor: Mark Fišel

Emergence of large language models (LLMs) has revolutionized natural language processing. However, with current training methodology high-quality LLMs can only be created for languages with abundant text datasets, which leaves out a significant portion of the world’s languages that do not have such resources. Moreover, development of LLMs for new languages also depends on the existence of benchmarking test sets for these languages; their creation, however is very expensive and depends on the availability of speakers of these languages, further hindering massive efforts on supporting low-resource language families. In this project we will focus on LLM development and evaluation for Finno-Ugric languages, a family that includes mid-resource languages (Estonian, Finnish, Hungarian), low-resource languages (Sami languages, Komi, etc.) and extremely low-resource languages (Livonian, Votic, etc.).
A prospective candidate should have a strong background in NLP and math with experience of text processing as well as training and evaluating neural networks.

Supervisor: Mark Fišel

Automatic text correction is a field of NLP and AI that particularly suffers from scarce data for all languages except English. At the same time it is a vital of digital support of languages, both for native speakers and language learners. This project will tackle the challenge of developing text correction models for low-resource languages including Estonian, with the help of large language models. We will approach it both from the angle of distilling abilities of closed commercial solutions (like GPT4) as well as fine-tuning open-source language models to perform tasks in automatic text correction, including grammatical error correction and dialect normalization.
A prospective candidate should have a strong background in NLP, experience of text data processing, prompting as well as training and evaluating neural networks.

Supervisor: Dirk Oliver Jim Theis

Quantum computing is gradually entering the era of fault-tolerance, a transition that significantly amplifies the complexity of enabling software, encompassing both firmware and design tools such as compilers. Addressing parallel decoding of error syndromes, code-specific quantum logical gate primitives, and logical qubits in various shapes and sizes demands attention at both the design and verification levels.

Within the Horizon-Europe funded project OpenSuperQPlus, the University of Tartu is contributing to the development of compiler and verification methodology and software. This initiative includes two PhD projects, each focusing on one of these areas.

The PhD-project on compilers aims to create intermediate representations for fault-tolerant quantum computing using the Multi-Level Intermediate Representations (LLVM-MLIR) framework. Simultaneously, the verification PhD-project will commence with coarse-grained testing based on the stabilizer-formalism and ZXH-calculus, progressively advancing towards effective unit-testing methodologies.

The projects will contribute to transferring of cutting edge quantum computing design skills into the Estonian workforce.

Supervisor:Radwa Mohamed El Emam El Shawi

In recent years, the confluence of artificial intelligence and environmental
sustainability has garnered significant attention. While AutoML has proven
transformative in making machine learning accessible to non-experts, the
environmental consequences of these automated processes have been overlooked.
Traditional AutoML methods tend to prioritize performance gains without due
consideration for the substantial resource-intensive computations and their associated
carbon footprint. The primary goal of my doctoral thesis is to tackle the increasing
environmental impact of automated machine learning (AutoML) processes by
pioneering innovative approaches within the field of green AutoML. As machine
learning models become more sophisticated and ubiquitous, the computational
demands for their development have raised concerns about their carbon footprint and
overall environmental sustainability. In response to this, my research is dedicated to
addressing the intersection of artificial intelligence and environmental responsibility.

Industrial doctorate, junior researcher

Supervisors: Kallol Roy, Madis Kiisk( Gscan OÜ)

Cosmic Ray Tomography (CRT), or Muon Tomography, uses cosmic-ray muons for 3D imaging to identify hidden objects, offering potential benefits in sectors like airport security, infrastructure inspection, and healthcare. This technique utilizes muon scattering data to differentiate materials, aiming to be a viable alternative to X-ray technology for 3D imaging and material classification. However, its practical implementation requires overcoming challenges related to system performance, including scanning time, tracking accuracy, and cost-effectiveness. This project introduces an AI solution based on deep learning (DL) for inverse image reconstruction and material classification in CRT. It leverages differentiable programming (DP) to build flexible ML models trained on muon tracking data. These models dynamically adapt to input data and detector configurations, outperforming traditional methods by learning modelling directly from training data, leading to innovative solutions in complex, non-intuitive solution spaces.

Supervisors: Veljo Kisand, Kristel Panksep, Leho Tedersoo

Freshwater and estuarine ecosystems face threats from climate change and human activities. Yet, a crucial group of players in their health, aquatic fungi (AF), remain largely unknown. Despite their essential roles in nutrient cycling and decomposition , only a fraction of the estimated 20,000 AF species are identified. This knowledge gap hinders effective conservation efforts. This project aims to unlock the secrets of AF across Europe. Using cutting-edge DNA analysis, we will unveil their hidden diversity, explore their habitat preferences, and understand their critical roles in ecosystem health. By bridging this knowledge gap, we empower researchers and conservationists to develop effective strategies for protecting these vital organisms and the ecosystems they sustain. The project holds significant potential from the perspective of microbial ecology, as understanding the relative microniche and host preferences, including pathogenesis, of fungi provides valuable insights into the broadness of fungal niches. This knowledge can be instrumental in shaping conservation strategies and leveraging the diverse roles fungi play in ecosystems.

Supervisors: Hanna Hõrak, Ebe Merilo

Stomata are small pores in leaves that mediate carbon dioxide uptake for photosynthesis and water loss via transpiration, while allowing entry of plant pathogens and air pollutants. Stomatal numbers and apertures affect plant gasexchange, growth, disease resistance and yield. Most plants have stomata only in the lower (abaxial) leaf surface, whereas some amphistomatous plant species, including the model plant Arabidopsis (Arabidopsis thaliana) also develop stomata in the upper (adaxial) leaf surface. Very little is known of how adaxial stomata are formed and what is their role in plant gas-exchange, growth, yield and disease resistance. The aim of the project is to understand, how adaxial stomata are formed in Arabidopsis and how the ratio of adaxial and abaxial stomatal densities (stomatal ratio) affects plant gas-exchange, growth and stress resistance. During the project, plant lines with altered adaxial stomatal densities will be isolated, the role of the phytohormone abscisic acid in the regulation of stomatal ratio will be addressed, and gas-exchange and growth traits will be characterised under normal and stress conditions in plant lines with modified stomatal ratios. The project will result in an improved understanding of the role of adaxial stomata in plant physiology and their importance for agronomically important traits.

Supervisors: Dr. Kuno Kasak, Dr. Evelyn Uuemaa

The rising levels of carbon dioxide and methane in the atmosphere are driving climate change, leading to global temperature increases and impacting ecosystems. Wetlands, crucial carbon reservoirs, pose challenges for studying their carbon cycle due to their complexity. Understanding methane emissions from wetlands is particularly challenging due to episodic fluctuations influenced by various factors. Combining different observation methods offers a chance to study methane flux variations better. This doctoral project aims to use data from Estonian wetlands and international databases to study methane flux heterogeneity and drivers, ultimately developing a model to improve flux estimations. The specific objectives of this PhD project is a) to analyze hot spots and hot moments of GHG emission in the wetlands in different climate zones and driving mechanisms; b) to analyze the spatial and temporal heterogeneity of GHG fluxes in various wetlands and c) upscale the heterogeneity to global wetlands. The core research will be carried out in three wetland ecosystems in Estonia and five wetlands in the SacramentoSan Joaquin Delta in California with strong collaboration with other international partners in Europe, Asia, and the United States.

Supervisors: Michael Dannemann, Irene Gallego Romero, Kaia Palm

Humans have been exposed to a large variety of pathogens throughout their existence. Being able to adapt to them was a matter of life and death. Genetic variants that helped to fight pathogens remained in the human gene pool. Today, we still carry some of these genetic modifications in our DNA - modifications that impact human health. 

This project seeks to delve into the impact of past episodes of natural selection and admixture with ancient modern human populations and archaic humans on the composition of human genomes in response to pathogen exposure. To achieve this, the project will utilize immune datasets capable of establishing a link between pathogens and immune response phenotypes. Through the analysis of these datasets, the project aims to pinpoint specific loci associated with pathogen-specific immune responses. Investigating the evolutionary history of these loci will unveil the impact of past pathogen exposure on shaping the genomes of present-day populations. 

The findings of this research are expected to yield novel insights into the relationship between pathogens and disease-associated variants in present-day individuals, potentially informing the development of innovative therapeutic strategies.

Supervisors:Andres Metspalu, Mart Kals

Circulating tumor DNA (ctDNA) is a biomarker that could be used for early detection of cancer. Oxford Nanopore Technologies has developed Oxford Nanopore sequencing (ONT) that enables the sequencing of DNA even in the presence of only one copy of the DNA molecule. In addition, a protocol has been recently developed that enables the analysis of ctDNA methylation patterns with ONT. In this work we will use clinical cancer patients’ liquid biopsies (blood plasma) and ONT to prove whether it’s possible to detect cancer in very early stages, using ctDNA methylation analysis, and in the case of cancer how many months/years in advance is it possible to detect ctDNA in blood plasma. Such a method would facilitate early clinical intervention. Furthermore, we’ll detect the tissue of origin of the ctDNA using its methylation pattern which in turn would allow for testing of multiple cancers at the same time with just one singular method.

The phenotypes in question are breast cancer, prostate cancer, colorectal cancer, and melanoma. The study sample would consist of clinical cancer patients from the Tartu University Hospital and gene donors from the Estonian Biobank.

Supervisors: Mait Metspalu,Ester Oras,Antonio de Dios Martinez,Remi Philippe Barbieri

Recent advancements in ancient DNA (aDNA) research, facilitated by improved laboratory methods and sequencing technology, have revolutionized our understanding of prehistory. By accurately dating and geolocating samples, aDNA enables comprehensive studies of past genetic diversity, addressing key questions in human demography. However, the scarcity of human remains in certain periods and regions poses challenges. To overcome this, birch tar, a novel source of aDNA and proteins, is being explored. Birch tar, commonly used in the Stone Age, contains human and microbial DNA, providing insights into ancient diets and pathogens. Metagenomic analyses of chewed birch tars reveal information about the human oral microbiome and potential infections. Additionally, DNA and proteins from food particles in birch tar offer direct evidence of ancient diets, complementing traditional methods with higher resolution proteomics. This interdisciplinary approach promises to enhance our understanding of ancient populations and their lifestyles.

Supervisors: Priit Palta, Urmo Võsa, Jaanika Kronberg, Erik Abner

This PhD project aims to explore the genetic variation and blood metabolite associations as drivers of predisposition and progression of highly heritable common diseases. It will leverage data, tools, and results from PRG1291 and other ongoing or recently finished projects like CVDLINK, DISCERN, EXPANSE, and FinnGen collaboration. The project is structured into three articles: 
Article 1 focuses on the characterisation of genetic, metabolomic, and comorbidity-based disease networks. It plans to calculate and analyse networks using genetic correlations between diseases defined from GWAS, metabolite-disease association analysis and comorbidity scores. The goal is to identify disease clusters with possible shared underlying mechanisms, which will be further analysed.

Article 2 aims to develop genetic and metabolomic risk scores for a set of diseases. The diseases to be studied will be defined based on data from previous analyses. The models will be trained and validated on individuals of the Estonian Biobank, with potential replication using UKBB data.

Article 3 focuses on the study of the progression of a specific group of diseases. The main focus in the context of the development and progression of cardiovascular diseases (CVD) is on metabolic syndrome, which is currently strongly underdiagnosed. As understanding the environmental, genetic, and metabolite patterns at the time of the first appearance of symptoms is important for early screening and understanding disease development (especially since the pathological processes of many chronic diseases can start decades before symptoms emerge), the aim of this work is to identify genetic and/or metabolite patterns for the development and progression of CVD.

Supervisors: Burak Yelmen, Luca Pagani, Lili Milani

In the past two decades, genome-wide association studies (GWAS) have been the fundamental approach for understanding the genetic background of complex phenotypes with considerable success, resulting in the discovery of thousands of associated variants for hundreds of different phenotypes. However, conventional GWAS have certain limitations: (i) detection of small effects requires large datasets and (ii) model-based methodologies with simplified assumptions cannot reliably capture complex genomic landscapes affecting the observed variation of quantitative traits. Furthermore, converting association signals to causal explanations is non-trivial without functional analyses even with the most recent statistical methods, mainly due to high correlation between genomic loci. These issues result in polygenic risk scores obtained from GWAS summary statistics to have low predictive power for many quantitative traits. To address these challenges, this project aims to develop machine learning models such as artificial neural networks which have become state-of-the art for many tasks in multiple domains with high-dimensional complex data structures. These models will be trained to predict human quantitative traits from genotype data and potentially become the basis for better genetic risk prediction tools over conventional approaches for future clinical applications.

Supervisors:Uku Vainik, Priit Palta

Obesity is a heritable chronic condition, costing 2% GDP worldwide. Many obesity treatments - behavioural, pharmacological, and surgical have been developed. Intriguingly, people’s responses to treatments vary widely - some may lose a lot of weight whereas others may even gain weight! To predict such weight change variability, close to 200 measures have been proposed in the past. Still, there are too many measures to be used as predictors or intervention targets. We will use genomic causal inferences analysis to establish, which of the 200 measures could be causally linked with weight loss success. If we succeed, we will set a new standard for the behavioural health sciences allowing for quicker discovery of intervention targets.

Supervisors: Evelyn Uuemaa, Aveliina Helm, Alexander Kmoch

Habitat loss, fragmentation and degradation are increasing threats to global biodiversity, underlining the critical need for more efficient biodiversity assessment and monitoring. Ecosystem structure, a key determinant of habitat quality and a relevant indicator of biodiversity at local and regional scales, highlights the necessity of understanding the relationships between species richenss, ecosystem condition and habitat structure. To achieve the necessary resolution and accuracy for detailed ecosystem three-dimensional (3D) structure, LiDAR stands out as the most direct and precise method. Despite this, the full potential of combining active (LiDAR, radar) and passive remote sensing remains under-utilized. Moreover, multi-temporal (seasonal) feature sets, consisting of numerous combinations of spectral bands, can hold a potential to predict compositional vegetation classes.

The PhD project aims to 1) explore the use of high-resolution remote sensing indices for estimating functional and structural biodiversity in Estonia, 2) develop of a spatial machine learning model for nationwide biodiversity prediction. Utilizing LiDAR data from the Estonian Land Board, the project will generate detailed ecological descriptors for the entire country, including vegetation height, structure, density, and topographic features. Sentinel-1 and Sentinel-2 data will contribute various indices, and a multi-temporal aspect will be captured through principal component analysis. These metrics will serve as covariates for the machine learning model, establishing connections between remote sensing indices and biodiversity indicators derived from ground measurement which will enable modelling biodiversity condition at national level.

Supervisors:Tiit Tammaru, Kadi Kalm, Kerli Müürisepp, Veronika Kalmus

The overarching goal of the PhD project is to enhance our understanding of the changes in the spatial structures of inequality in digitally transforming societies by focusing on changes in residential segregation. Research on residential segregation has reached a turning point. It needs new perspectives and approaches since an important share of activities has shifted to the digital space, which in turn brings along changes in the physical space. The residential preferences of high-skilled professionals are critical in shaping patterns of residential segregation, but how these preferences change during digital transition is unclear. The proposed project will address this research gap through the study of the Estonian capital city, Tallinn. Estonia is well known for its cutting-edge digital solutions for businesses and the public sector. We will apply a comprehensive longitudinal and geospatial research design that will draw on and harmonised microdata from censuses and registers from the Estonian Longitudinal Database.

Supervisor: Ivika Ostonen-Märtin

The most significant uncertainty regarding the impact of climate change on terrestrial ecosystems lies belowground. Plant roots serve as the primary channel for carbon entry into the soil, shaping vital interactions within belowground biodiversity. Among the components of tree root systems, fine roots assume a critical role, given their dynamic and short-lived nature. This unique trait enables them to adapt to changing soil conditions and influences various mechanisms associated with carbon storage. This PhD project aims to enhance our understanding of belowground carbon cycle dynamics, in relation to various complex responses of plants to changing climatic conditions, mainly via root functional, structural, and chemical parameters. Focusing on the C accumulation driven by fine roots and their complex interactions in the rhizosphere, involving host trees, its roots associated ectomycorrhizal (EcM) fungi and other microbiota in the soil. Commencing with the enhancement of our knowledge about crucial processes belowground, influencing terrestrial ecosystems, this projects also aims to evaluate the role of fine roots related indicators in next-generation soil monitoring programmes. The goal is to address and optimize the fragmented information on roots in European soil monitoring systems, while advocating for the integration of root data, as an essential foundation for successful soil policy and informed land management practices. This integration aims to improve soil health assessments and contribute to the development of a harmonized soil monitoring framework, fostering sustainable forest management practices.

Supervisors: Veronika Mooses, Age Poom

The aim of the PhD project is to explore the equity issues in low-carbon mobility transition process. Research has shown that sustainable mobility options using novel technologies such as bike-share and e-scooter infrastructure has been mostly concentrated in the wealthy city-centres and equity issues are generally overlooked in policy-making, which limits the accessibility to low-carbon mobility modes of certain population groups and can increase existing inequalities. In this project, equity is tackled from the spatial and population perspective in Estonian and European cities. Spatial perspective deals with distributive and procedural justice, i.e. how sustainable mobility solutions are spatially planned, allocated, and accessed. Population perspective deals with questions of on what conditions accessibility to sustainable travel modes is translated into actual behaviour, and how to ensure the right to be mobile for vulnerable groups (children, elderly, people with disabilities) in low-carbon mobility transition. The role of universities – as large knowledge-intensive organisations triggering innovation – in sustainable mobility transition is also tackled. The goals of the PhD project call for a mixed methodology approach of using both quantitative and qualitative data and methods.

Supervisors: Mikk Espenberg, Ülo Mander

Nitrogen (N) plays a critical role in supporting global life and ecosystem functioning, with implications for climate regulation. However, human activities and climate change have disrupted N cycle processes, leading to imbalances and losses, including eutrophication and N2O emissions. Peatland ecosystems, closely linked to biogeochemical cycles, highlight the significance of studying N due to its role as a limiting nutrient. Land use changes can shift peatlands from N sinks to sources, impacting global climate change. Understanding the pivotal role of N2O, a potent greenhouse gas, is crucial for mitigating its impact. Although key players in peatland soils producing N2O have been identified to some extent, their roles remain understudied. A comprehensive understanding of N cycle processes and the microbiome is essential for predicting peatlands' response to land-use changes and their influence on global climate warming.

This PhD project aims to (1) integrate a comprehensive database and incorporating new LiWeFor and PeatlandN2O (ERC) project sites on peatlands' microbiome; (2) develope a bioinformatical pipeline for studying the microbial N cycle; and (3) predict the land-use effect on the microbial N cycle, along with concurrent N2O fluxes in global peatlands.

Supervisors: Ülo Mander, Margit Aun

Among greenhouse gases with a substantial anthropogenic component to their emissions, N2O has overall the third strongest global climate forcing effect after CO2 and CH4 (IPCC 2021). While N2O is mostly non-reactive in the troposphere, it is currently the primary ozone depleting  gas in the stratosphere which is not regulated by the Montréal protocol (Ravishankara et al 2009; Stolarski et al 2015; Tian et al 2020). The instigating articles for this thesis project are Ricaud et al (2009), where very high N2O emissions were derived from satellite measurements of N2O in the total air column above tropical rainforests, and Ball et al (2018), which showed a continuous decline in lower stratospheric O3 over tropical latitudes, possibly being caused by increasing N2O concentration in stratosphere. The presented aspects make N2O research up-to-date, emissions have increased in recent decades and the long life of N2O has led to an increase in the importance of the climate warming effect compared to other greenhouse gases (IPCC 2021). The objectives of the doctoral thesis are: (1) to analyze N2O and O3 spectral image measurements in the stratosphere during the period 2005-2021 derived from the NASA Aura satellite MLS instrument, (2) to analyze relationship between the N2O and O3 concentration in stratosphere and the altitude, season and geographic location; (3) to analyze the long-term N2O trend in the stratosphere, (4) to analyze the N2O fluxes in the troposphere measured by the IASI instrument of the MetOP and EUMETSAT satellites; (5) to relate tropospheric N2O fluxes and stratospheric N2O concentrations to ground-based measurements. The data used in the work come from publicly accessible databases: NASA Earthdata EOSDIS, EUMETSAT IASI, EL EDGAR, as well as from the database of N2O emission and corresponding environmental factors of global organic soils collected by the working group. For data processing python and R programming languages will be used. Project results will help clarifying N2O emission spatial and temporal patterns in wetlands. They will serve as a basis for land use optimization to mitigate N2O emissions.

Supervisors: Evelyn Uuemaa, Holger Virro 

The Common Agricultural Policy (CAP) was established by the European Union in 1962 to support sustainable farming practices and enhance agricultural productivity. One of the measures within CAP is greening, which involves identifying landscape features like vegetation patches and forest islands on agricultural fields. In Estonia, the Agricultural Registers and Information Board manually adds these features to a database based on farmers’ applications, a time-consuming process unsuitable for large-scale mapping. To address this, the PhD project aims to develop a deep learning (DL) methodology for detecting landscape elements (stone walls, hedgerows, small forest patches) on Estonian agricultural fields at a national scale. The project utilizes freely available remote sensing data, open-source software, and DL models based on convolutional neural networks (CNNs), such as the U-Net architecture. Training data for the models are generated by digitizing landscape elements from orthophotos provided by the Estonian Land Board, with potential integration of elevation data from Lidar points as a secondary input layer. The PyTorch Python package and TorchGeo library will be employed for DL model development and effective national-scale application. The developed methodology is applicable in other regions in the world where similar landscape features exist.

Supervisors: Ülo Mander, Lulie Melling, Mikk Espenberg, Jaan Pärn

Tropical peatlands are globally important carbon and nitrogen stocks. Cycling between these and atmospheric carbon and nitrogen is driving climate change and is the basis of life in tropical peatlands. The University of Tartu has a unique database of tropical peat samples with their microbiome analysed for CH4 (methane) and N (nitrogen) cycle functional genes. Molecular composition of dissolved organic matter in a representative subsample of the same soils has been analysed by high-resolution mass spectrometry. According to the analysis, molecules have been empirically categorized between compound groups ranging from proteins to aromatic. Isotopic composition (15N and 13C) of the soil samples will be analysed during the doctoral study. The doctoral thesis will associate these high-resolution biogeochemical fingerprints with both field measurements and annual model estimates of CO2, CH4, N2O and N2 gas fluxes to identify carbon and nitrogen cycling pathways. This should conceptually advance our understanding of tropical peatland biogeochemistry from the integrated isotope, microbiome and reactomics analysis. This will serve as a basis for the protection and sustainable management of tropical peatlands.

Supervisors: Mikk Espenberg, Lulie Melling, Ramita Khanongnuch

The importance of understanding microbially mediated methane (CH4) processes in peat is especially important, particularly in the context of peatland management approaches. The majority of atmospheric CH4 originates from biogenic sources, produced primarily by methanogenic archaea in anoxic conditions. This PhD project addresses the uncertainty surrounding the mechanisms regulating CH4 fluxes and emphasizes the need for specific considerations in tropics. The net CH4 emissions result from a complex interplay between CH4-producing and consuming microorganisms.

This PhD project aims to investigate the microbial methane CH4 cycle in tropical peatland soils, focusing on describing, explaining, and predicting CH4 dynamics. The specific objectives are to (1) identify and quantify key microbes as indicators for CH4 production and consumption using advanced DNA- and RNA-based methods, (2) explore the spatial distribution of these microbes within the peat depth and spatial heterogeneity across tropical peatlands under varying conditions, and (3) examine the role of different CH4 oxidation processes, including aerobic and anaerobic pathways, in tropical peatlands.

Supervisor: Oive Tinn

Although the presence of distinctive, but rare fossil chelicerates in Estonia has been known internationally from the 19th century, and has been referred to in a fair number of research papers, no detailed analyzes on chelicerate fauna has ever been done in Estonia. The most comprehensive compendium of the fossil record of Estonia does not treat the group at all, another includes a brief summary only. Except for a few overviews there is no detailed analysis of chelicerate facies distribution, stratigraphy, palaeogeography and palaeoclimate in the context of the Baltic Palaeobasin. 

Potential research papers:

I Chelicerate fossils from the Kalana Lagerstätte. This unique Silurian-age locality has been known for its exceptionally preserved algal fossils, but has also yielded rich faunal community, comprising crinoids, vertebrates, etc. As the Lagerstätte comprises extraordinary biota which has not been preserved elsewhere, the investigation of its chelicerate fauna is critical for the reconstruction of the marginal-marine ecosystems in the early Silurian. 

II Distribution and palaeoecology of chelicerates in the Silurian of Estonia. This sub-project includes a major revision of chelicerate taxa, palaeocommunity analysis incorporating total biota, and detailed stratigraphic and facies pattern analyses of chelicerate taxa. The work involves study of Estonian fossil collections (incl. Patten’s collection now in AMNH) and fieldworks in fossil localities. 

III Chelicerate distribution in the Baltic Palaeobasin. In the Silurian of the Baltic Palaeobasin, the best chelicerate fossil record has been recorded in Estonia. Additionally, chelicerate fossils have been reported from Sweden and Norway. The main objective is the analyses of chelicerate facies distribution in larger tectonic, palaeogeographical, palaeocological and palaeoclimate context.

Supervisors: Marko Kohv, Martin Liira

Objective: This project aims to analyse and mitigate the impacts of climate change on coastal areas, specifically addressing coastal hazards. The focus is on employing advanced unmanned monitoring technologies and developing predictive models for better risk management and adaptation strategies in coastal zones.

Methodology: The research integrates geology, geomorphology, and remote sensing across several phases. The initial phase involves analyzing existing data on coastal erosion. Subsequent phases include deploying unmanned technologies (UAVs and USVs) to collect detailed geomorphological data and intercalibrating this with traditional monitoring methods. The final phase concentrates on coastal hazard modelling and scenario development.

Collaboration: A key partnership with the Estonian Geological Survey (EGT) will facilitate the integration of new and historical monitoring data, crucial for understanding the long-term impacts of climate change.

Innovation: The project leverages advanced monitoring technologies to fill data gaps in shallow water zones and enhances risk prediction capabilities by developing advanced predictive models.

Outcomes: Expected outcomes include a refined understanding of the relationship between coastal features and erosion risks, validated new monitoring methods, and enhanced models for identifying coastal risk hotspots. These insights will contribute to the development of nature-based solutions for coastal risk management, adapting effectively to the challenges posed by climate change.

Supervisors: Kalle Kirsimäe, Timmu Kreitsmann

In past decades, REEs have emerged as essential commodities for advanced technologies and the REEs are widely used in fields ranging from metallurgy to energetics. The REEs have become vital for green technologies and particularly for the ongoing transition to sustainable energetics. Demand for REE is estimated to increase by ca. five times in the next 20 years. The majority of the REEs are found in carbonatite deposits enriched in LREE, and the HREE are primarily concentrated in alkaline–peralkaline deposits. However, the late-phase granites, particularly their weathering products (i.e., ion-adsorption clays), are REE resources of growing importance. This project studies the REE metallogenetic potential of the rapakivi granite batholiths and their weathering products buried under sedimentary cover in Estonian and western Latvia is unknown and warrants a targeted study.

Supervisors: Kalle Kirsimäe, Peeter Paaver

During Fe-sulphide formation, various elements are incorporated at trace to minor levels via substitution into the lattice or as nano- to micro-scale inclusions. Common elements include Co, Ni, As, Se, Zn, Cu, Au, Ag, Te, Pb, Sb, Bi, Tl, V, Mn, PGEs, and Hg. As, Co, Ni, and Se substitute for Fe or S, while Cu, Zn, Te, Pb, Bi, Ag, and Au occur mainly as nanoparticles or micro-inclusions. The trace-element composition reflects environmental conditions, aiding ore exploration. In the Estonian crystalline basement, sulfide mineralization is mainly pyrrhotite with lesser pyrite and chalcopyrite. Limited testing on Jõhvi magnetite ores shows potential for base and precious metal exploration. This PhD project aims to characterize sulfide mineralization in the Estonian crystalline basement, determine trace-element distribution using in-situ methods, and reconstruct the sulfide formation's fluid activity and environmental conditions.

Supervisors: Riho Mõtlep, Peeter Somelar

In Ida-Virumaa, hudreds of million tons of various wastes from the oil shale industry are concentrated, which are practically not used, but which represent a potentially valuable secondary raw material for various industries. Such waste deposits have a diverse origin, composition and development history, which complicates their use. Systematic studies of the composition and heterogeneity of waste make it possible to identify different possibilities of use in the chemical industry, the building materials industry, the fertilizer industry and as aggregates. The general goal of the project is to study the composition of Ida-Virumaa's solid waste through a systematic approach and to evaluate the best opportunities for its valorization. For this purpose, an overview of the solid waste of Ida-Virumaa will be undertaken, studies of the composition and heterogeneity of solid waste are carried out, and possible existing solutions for valorizing waste in both the chemical industry and the building materials industry are found.

Supervisors: Vambola Kisand, Alexander Vanetsev, Angela Ivask

The PhD project will focus on the development of antimicrobial coatings that act efficiently in dry and semi-dry conditions. We plan the preparation and testing of stable antimicrobial coatings based on two kinds of promising antimicrobial nitrogen-containing compounds, alkylated PEI (polyethylenimines) or N-halamines. The work will include synthesis of antimicrobial polymeric salts based on alkylated PEI or N-halamines, developing/selecting the best covering procedure (spray, paint, immersion), developing the methods of covalent attachment (with or without chemical modification of antimicrobial compounds). Composition, uniformity and thickness of obtained coatings will be studied. Testing of the stability of coatings for wear and tear and antimicrobial tests will be performed. Additionally, the PhD project will cover combinations of known or proposed antimicrobial compounds to create surfaces with synergistic antimicrobial effect. Combinations between the most efficient alkylated PEI and N-halamine compounds together with several inorganic antimicrobial compounds (including TiO2, ZnO, copper and silver metal nanoparticles) will be prepared and used to create surfaces. Those selected compounds act via different modes of action (photocatalytic effect or metal toxicity combined with cell membrane rupture effect by alkylated PEI or halamines) and are therefore expected to exhibit broad antimicrobial effect. The latter would ensure significant and prompt microbicidal effect and decreased possibility for surviving microbial cells, including potentially resistant microbes. Three scientific publications are foreseen within the PhD project. The present application relates to TEM-TA55 project “Antimicrobial synergy-driven surface coatings - innovative solutions in healthcare environment”.

Supervisor(s): Kaupo Kukli, Lauri Aarik

The project will seek thin films of metal oxides simultaneously possessing optical transparency, electronic conductivity, hardness and elasticity. Oxides of different metals, such as gallium, niobium, titanium, hafnium, zirconium, aluminum, as well as their combinations will be synthesized at low temperatures and thoroughly analysed. Advanced materials layers will be prepared, consisting of ternary compounds formed after doping and mixing host solids with foreign elements. During the 4-year project, the PhD student will have access to state-of-art laboratories and equipment used for material studies. The student will exploit the atomic layer deposition method for the synthesis of abovementioned materials layers and use the contemporary electrical probe station for electrical measurements for their evaluation. Additionally, the student will have the possibility to work on different characterisation tools, while studying the elemental composition, structure and optical properties of thin films. All this will be used with the aim at the determination of the thin film structures and material combinations with chemical and physical properties allowing applications in nanoelectronics and optics. Concurrently, the main result is the graduation of highly qualified and skilled young researcher with doctoral degree, able to contribute to the further development of science and technology.

Supervisor(s): Taivo Jõgiaas

The use of secondary raw materials for the production of technologically important materials is investigated within the framework of this doctoral thesis. The work examines the mechanical properties of the manufactured materials and the technological methods required for their production. For example, silicate bricks left over from the demolition of buildings and carbon from plastic or wood waste are used as raw materials.

Supervisor(s): Alvo Aabloo, Nadežda Kongi

This doctoral project addresses the contemporary trends and obstacles in Electrochemical Additive Manufacturing (ECAM). It aims to investigate the potential enhancements achievable by integrating low-temperature eutectic molten salts, particularly in speeding up printing processes and improving material properties within ECAM. The project is structured to explore innovative materials and techniques to overcome challenges such as slow deposition rates and the fabrication of small microstructures. This doctoral project also aims to investigate specific applications of ECAM-produced micro and nanostructures across industries. By combining theoretical analysis with experimental application, the research aims to contribute to developing ECAM technology, particularly in improving the quality of materials produced and their suitability for use in various sectors such as medical devices, robotics, and electronics.

Supervisor(s): Tarmo Tamm

The trend of the green transition drives the search for increasingly efficient and environmentally friendly materials and solutions, among those the valorization of waste sources is of particular interest. The use of lignocellulose from secondary sources in the reinforcement of masonry materials makes it possible to hit several targets at once: significantly improve the properties of the material, make it more environmentally friendly, as well as remove carbon from circulation. The aim of this project is to investigate the use of lignocellulosic derivatives for the reinforcement of masonry, including autoclaved aerated concrete produced in Estonia. The first order is to find an answer to the question of whether cellulose derivatives, such as the much-studied hydroxyethyl methylcellulose (HEMC), are involved in chemical processes that proceed during autoclaving. Understanding the corresponding mechanism allows the design of highly efficient aerated concretes with improved properties. There is also great interest in employing additive manufacturing for producing concrete constructs. The problem here is both the fusion of different layers upon extrusion and the execution of the reinforcement necessary to achieve strength. The current project is looking for ways to create an initially soft but (chemically) hardening rebar based on lignocellulosic derivatives. The HEMC additive should make it possible to optimize the rheological properties and the strength of the resulting material. In order to make any of the results industrially applicable, it is crucial to study the effects of upscaling on the properties of such composites, which have been virtually unexplored so far. The results of this project will support the acceleration of the circular economy by prioritizing effective environmentally friendly approaches and resource efficiency, cooperation with Estonian companies is also envisaged.

Supervisors:  Kristo Väljako, Ülo Reimaa

The central concept the project tackles is situations where an object is equipped with the means to induce change in another object. For example, like inputs would act on an automaton or symmetries would act on a space. Such a means of acting is called an action. Using the automaton analogy, an action is partial if the machine can be in a state such that for some inputs the behaviour is undefined. In the context of mathematical objects, undefined behaviour is typically avoided, but partial operations can sometimes be either unavoidable or the simplest means of abstraction. Variations of the concept of an action can be found everywhere in mathematics, and while not all of them fit under the same umbrella, subsets of them do admit common generalizations. This project aims to do the same for partial actions—to gather a plethora of the more common and well-behaved instances of the notion under a unified point of view. Category theory is well-suited as the setting of such a generalization, since categories are the context in which actions are typically considered, and there are well-established ways for equipping a category with a notion of partiality. The minimal class of examples the project aims to tackle includes partial actions of the more common acting objects, such as groups, monoids and semigroups, and their various enriched counterparts, such as Hopf algebras, rings and Banach algebras, for which there already is a categorical approach to total actions. One of the benefits of a unified approach is that it helps one see through the theory more clearly and avoid the mysterious intricacies that may appear when restricted to one very particular setting. A categorical approach helped explain some of the peculiarities that appeared when studying total actions of semigroups. The study of partial semigroup actions has revealed even more intricacies, and part of the project is to explain why these intricacies appear from a more abstract point of view.

Supervisors: Arvet Pedas, Mikk Vikerpuur

Fractional (non-integer order) derivatives and differential equations involving such derivatives have nowadays become a subject of intense study, due to their broad application in the modelling of real-life problems. It has become evident, however, that many processes can not be effectively modelled using only constant-order fractional differential equations. A novel generalization for such problems are variable-order fractional differential equations, where the order of the fractional derivative of the unknown function changes continuously with respect to time. A comprehensive theoretical and numerical study of such equations is the main objective of the project. This requires a detailed examination of existence and regularity of solutions of the underlying problem, since even solutions of constant-order fractional differential equations are typically non-smooth. Based on the obtained regularity information on the exact solutions of variable-order fractional differential equations, we plan to construct and justify high-order methods for their numerical solution. The final step in the project is the validation of our proposed methods by detailed numerical experiments.

Supervisors: Viktor Abramov, Olga Liivapuu

The research subject of this project lies at the intersection of two areas of mathematics: algebra and differential geometry. The theory of Lie algebras, Lie superalgebras, and their generalizations plays an important role in modern differential geometry and theoretical physics. Recently, a new and promising direction of research, which can be called transposed Poisson algebras and their relation with Novikov-Poisson algebras and 3-Lie algebras, has appeared in this area. The goal of this project is to study a super-generalization of a transposed Poisson algebra: a transposed Poisson superalgebra. This goal includes finding identities for transposed Poisson superalgebras, constructing examples of transposed Poisson superalgebras, the classification of transposed Poisson superalgebras by means of ½-biredivations, their representations, and relation to 3-Lie superalgebras.

Supervisors: Krista Fischer, Lili Milani

One of the main tasks of this projects would be to investigate the implementation of the methodology of Target Trial Emulation (TTE) in a biobank context. It is well known that Randomized Controlled Trial (RCT) is a gold standard in study designs enabling causal conclusions in biomedicine. However, with the emergence of big databases, such as population-based cohorts with electronic health records, the question on estimability of causal effects in observational studies is getting more and more attention, especially as the conduct of an RCT often faces ethical and practical burdens. One way to minimize biases in estimation of causal effects from observational data is to use techniques that mimic the protocol of a hypothetical RCT as closely as possible. The process requires care in sample selection and as well as the use of various data analytic techniques (such as digital “cloning” and censoring of individuals) and certain causal modelling approaches. Biobanks (such as the Estonian Biobank) have the advantage of combining longitudinal health, treatment history and –omics data in large population-based cohort. Therefore, a TTE study on a biobank cohort would enable not only the estimation of effects of certain treatments, but also estimation of gene-treatment interactions. A second research area for this project is related to the effects of Polygenic Risk Score (PRS) in combination with mammography screening on the incidence of Breast Cancer as well as on BC mortality. Although there is convincing evidence that mammography screening reduces BC mortality, there is still some uncertainty regarding the magnitude of the effect and costeffectiveness.

Supervisor: Rainis Haller

This doctoral project explores the concept of plasticity in Banach spaces, focusing on the behaviour of non-expansive bijections. Plasticity, defined as the preservation of distances under such mappings, poses a fundamental question: Is the unit ball of every Banach space plastic? Through a systematic investigation, we aim to advance the understanding of plastic pair’s theory, examine weakened plasticity by considering continuous inverses, and analyse the implications of non-expansive bijections on homeomorphisms. Our research plan involves addressing these objectives sequentially, with each contributing to the submission of research papers. By shedding light on these fundamental aspects, our study seeks to deepen the understanding of Banach space structures and contribute significantly to mathematical theory.

Supervisors: Krista Fischer, Lili Milani, Maris Alver

Genetic variation, drug-drug interactions, pre-existing health conditions, age and sex significantly impact drug efficacy and can elevate the risk of adverse drug reactions. Research in these areas has thus far not accounted for the extent of comorbidities and polymedication present in real healthcare settings. Real-world data captured from electronic health records provides a unique opportunity to assess these effects on clinically relevant treatment outcomes and adverse events in a systematic manner. The aim of this doctoral project is to determine the influence of genetics in drug response by taking a broader approach towards pharmacogenomics. We plan to integrate data on genetic variation, physiology, pharmacokinetic profiles, co-medications, age, and sex to explore their interactions and effects on treatment outcomes extracted from real-world data of large-scale biobanks. For the analyses one must acknowledge that a statistical interaction is not necessarily indicative of an underlying biological interaction – thus, one has to aim for the estimation of causal interactions rather than simple interaction testing in the models. Population pharmacokinetic models are based on nonlinear mixed-effects models involving differential equations and involve considerable mathematical complexity. Therefore, a mathematical background is essential for a PhD student working on this interdisciplinary project. The results will allow us to test and develop highly innovative stratification and risk prediction models for pharmacogenomics. The research will result in new models for substantially improved therapy and advance our knowledge of the extent to which real-world data can be used to investigate effects and interactions of factors which cannot be captured in clinical trials.

Supervisor: Hedvig Tamman

Bacteria living in a stressful environment are mainly in a metabolically inactive and dormant state. This state is caused by an activated stringent response which results in accumulation of alarmone ppGpp and reprogramming of bacterial metabolism from growth to stress survival. These dormant bacteria survive stressful conditions much better. The viruses of bacteria (phages) can still often infect and proliferate in these dormant bacteria, which refers to their ability to overcome bacterial stringent response. Phages manipulate bacteria to exit the stressed condition and become metabolically active, thus sensitizing them to stresses. Learning from these phage mechanisms will help us to develop mechanisms that would allow to keep bacteria in a sensitive state, so that antimicrobial methods remain efficient. The current doctoral project aims to shed light onto the interactions between bacterial stringent response and phages. We plan to determine the different mechanisms with which the stringent response protects bacteria from phages and how phage-bacteria interactions differ for bacteria in stressed or relaxed state. This allows us to better understand the arms race between stringent response and phages and brings us closer to the knowledge of how to manipulate the bacterial stress response so that bacteria would remain sensitive to antibiotic treatments.

Supervisor: Tiina Tamm

All viruses, regardless of genome type and size, utilize the host cell’s translation machinery, specifically ribosomes and translation factors, for the synthesis of viral proteins. Certain viruses employ specialised RNA elements, called internal ribosome entry sites (IRESs) to recruit ribosomes directly to their RNA. The use of such internal initiation allows specifically decrease the cellular protein synthesis and increase the translation of viral proteins. The Cricket Paralysis Virus intergenic region (IGR) IRES facilitates translation of viral structural proteins. This project uses genetic, biochemical and structural approaches to clarify how the IGR IRES-mediated translation initiation is regulated during virus infection. The project focuses how viral protease alters the functionality of cellular translation machinery. The project results in a deeper understanding of virus-host interactions, which in turn allows the development of therapeutic strategies to effectively control the viral diseases.

Supervisor(s): Angela Ivask, Merilin Rosenberg

The proposed thesis project is aimed at understanding bacterial responses to multiple stressors on dry-use antimicrobial surfaces, particularly focusing on end-use scenarios. It addresses a critical knowledge gap in how bacteria adapt and potentially develop tolerance under dry conditions and in the presence of biocidal materials. This research is underscored by recent findings that demonstrate diminished efficacy of reputedly potent antimicrobial surfaces, such as copper and silver, under semi-dry conditions, highlighting the urgent need for better targeted passive antimicrobial strategies. The project intends to help fill the existing knowledge gap, to develop a comprehensive understanding of bacterial defense mechanisms and inform the design of effective antimicrobial surfaces, addressing application-relevant challenges. The successful applicant will work on the following: 

 Bacterial evolution experiments, data analysis and phenotype characterization to investigate evolution of tolerance and adaptive mutations under dry conditions, aiming to understand adaptive mechanisms and enhance test format development for biocide tolerance development on antimicrobial surfaces. 

 Comparative analyses of genetic backgrounds and antimicrobial susceptibility across standard and laboratory bacterial strains, to refine strain selection for efficacy testing and generalize findings to clinically relevant bacteria.

  Exploration of natural transformation events on antimicrobial surfaces, focusing on the design of test strains and DNA sources to understand non-conjugative horizontal gene transfer on antimicrobial surfaces.

Supervisors: Tõnis Org, Elina Aleksejeva and Andres Salumets

The uterus, an essential organ in female reproduction and embryo implantation, undergoes hormonal changes that impact its cellular and molecular composition. Understanding these dynamic changes is particularly crucial for improving the outcome of infertility treatments. Extracellular vesicles (EVs) that are present in uterine fluid (UF) likely play a significant role during the early stages of conception as EVs contain regulatory and metabolic cargo of endometrial tissue, targeting developing embryo. The potential of using UF-EVs samples as a semi-invasive alternative to traditional biopsies for endometrial receptivity testing and endometrial cancer diagnosis constitute a compelling and innovative research area in female reproductive (tract) health studies. Endometrial organoids (EOs) are in vitro models that recapitulate the endometrium's hormonal responses. Human blastoids, derived from embryonic stem cells, mimic early blastocyst development. Used alongside EOs, they offer promising models for exploring molecular pathways and EV functions in implantation, in interaction of blastoids with EOs. This thesis proposal seeks to elucidate the molecular mechanisms related to the function of UFderived EVs throughout the menstrual cycle, and their role in the peri-conception period during embryo implantation.

Supervisors: Kaspar Valgepea, Kurshedaktar Majibullah Shaikh

Gas fermentation provides a unique opportunity for the circular bioeconomy by enabling carbon recycling from gaseous waste feedstocks into value-added bioproducts using microbes. Acetogens are ideal biocatalysts for gas fermentation as they use gas (CO and/or CO2+H2) as their sole carbon and energy source. However, rational metabolic engineering of acetogen cell factories is hindered because understanding of genotype-phenotype relationships is minimal. This project aims to establish a new workflow for making large-scale arrayed CRISPRengineered microbial strain libraries and to engineer superior gas-fermenting cell factories through advanced systems-level understanding of acetogen metabolism. The project will also determine essential genes for autotrophic growth of the model-acetogen Clostridium autoethanogenum.

Supervisors: Margit Mutso and Reet Kurg

The cytoskeleton plays a pivotal role in cellular architecture, providing structural support, defining cell shape, and orchestrating the spatial arrangement of essential components such as the nucleus. This finely tuned structure is subject to constant regulation, undergoing cycles of degradation and reconstruction throughout the cell cycle. Key parts of the cytoskeleton include actin, tubulin and myosin, among others. Disruptions in the cytoskeleton often lead to various pathological conditions, including cancer and neurological disorders. The TRMT112 network comprises seven known methyltransferases: WBSCR22, N6AMT1, METTL5, THUMPD2, THUMPD3, TRMT11, and ALKBH8. None of these methyltransferases have been linked to the cellular cytoskeleton. However, our preliminary findings suggest that N6AMT1, WBSCR22, and TRMT112 itself interact with tubulins and/or are influencing the expression patterns of tubulin. The primary objective of the current doctoral study is to investigate the TRMT112 network proteins involvement in the regulation of the cellular cytoskeleton. The study will contribute to a more thorough understanding of their impact on cytoskeletal regulation and by delving into the intricacies of cytoskeletal dynamics in both cancerous and non-cancerous contexts, the study aims to shed light on potential therapeutic targets and mechanisms underlying disease progression.

Supervisor(s): Assoc. Prof. Taavi Lehto, Dr. Helena Sork, Dr. Tõnis Lehto

CRISPR-based gene editing systems have had enormous impact on both basic and translational science and collectively hold enormous therapeutic potential for the treatment of human disease. CRISPR-based therapeutic entities have limited bioavailability and their wider translation into therapeutic use requires the development of effective and safe drug delivery systems. Peptide-based drug delivery systems have lately been successfully used for the delivery of nucleic acid-based therapeutics, including CRISPR-based endonucleases. In this project, we will develop novel peptide carriers based on our lipidated delivery peptide platform for CRISPR systems utilizing both the ribonucleoprotein (RNP) and modified messenger RNA (mRNA) format. Peptide/CRISPR nanoformulations will be studied for efficacy and safety in various gene editing and therapeutic cell culture models. Furthermore, we will investigate peptide/CRISPR nanoformulation for their biodistribution, gene editing efficacy and therapeutic application in the mice models in vivo. Together these studies will shed light on the peptide-mediated delivery of CRISPR and RNA-based therapeutic modalities and lay foundation for future pre-clinical and clinical studies.

Supervisor: Professor Andres Merits

In their hosts, different alphaviruses cause encephalitis or fever, rash and arthritic symptoms. In infected cells, these viruses form membrane-bound replicase complexes consisting from viral nonstructural proteins (nsPs), RNA, host proteins and membranes. This project aims to study the molecular interaction between nsP4 and nsP1, nsP4 and nsP2, and nsP4 and RNA. Analysis of determinants responsible for alphavirus RNA replication will be used to select novel combinations of replicase components that can be used to develop next generation of self-amplifying (sa) RNAs potentially usable as vaccines. We aim to optimize the levels of expression of gene of interest, the duration of expression and select replicase components that will interact with host cell (including with innate immune system) at desired manner. The saRNA systems will be tested in vitro as well as in vivo. In addition, project involves characterization of chemicals interacting with virus replicase and suppressing its activity.

Supervisors: Mart Loog, Ilona Faustova

Rapid advancing of bioengineering field creates a broad array of technologies, ranging from the sustainable production of chemicals, biomaterials, and drugs to various medical innovations. It requires the development of a new synthetic cellular regulatory system that would help to design efficient cell factory producer strains. Approaches known and widely used so far largely concentrated on transcriptional regulation. However, in eukaryotic organisms, the most common and fastest method of signal transmission is through the phosphorylation of proteins by protein kinases. This project aims to establish a new approach to phosphorylation regulation using cyclin-dependent kinases. By employing engineered cyclins and two-input system, we plan to create synthetic phosphorylation circuits that can be efficiently applied to modulate protein degradation. We will design a set of protein fusion tags to encode specific phosphorylation patterns, allowing for adjustable control over signal strength. The developed toolbox of phosphorylation tags has the potential to be widely used in regulation of cell signaling networks for various future applications.

Supervisors: Lauri Vares, Livia Matt

The linear production and consumption of plastics today is unsustainable. It creates large amounts of mismanaged waste, pollution, and CO2 emissions. To fulfill the global climate targets, the plastic recycling rates should be considerably increased, and the production of virgin fossil-derived plastics should be gradually phased-out. The main challenges associated with physical recycling include deterioration in material properties after repeated recycles and the need separate and purify mixed or contaminated wastes. In this project new chemically recyclable polymers intended for demanding applications will be designed and synthesized from bioderived building blocks. In chemical recycling, the polymers are converted chemically back to monomers, which can be then repolymerized using the existing manufacturing process. In such an approach, the quality of recycled plastics will not decrease. To achieve the material characteristics required in demanding applications, various aromatic and cyclic structures will be incorporated into polymer structure. This research is part of the “Centre of Excellence in Circular Economy for Strategic Mineral and Carbon Resources", 2024-2030.

Supervisor: Velle Toll

Anthropogenic aerosols offset an uncertain fraction of global warming induced by greenhouse gases. Rapid future warming would be expected if aerosols currently exert a strong cooling effect on Earth's climate. This project evaluates the plausibility of strong present-day aerosol cooling through aerosol impacts on clouds. Aerosols could strongly cool the climate if, in addition to leading to more numerous and smaller cloud droplets, they made clouds thicker and more extensive. This project analyses a large dataset of ship tracks – linear cloud features polluted by shipping emissions that can be identified in the satellite images of clouds. Ship tracks serve as natural experiments of aerosol-cloud interactions, where the properties of aerosol-polluted clouds can be directly compared to the properties of nearby unpolluted clouds. This means that ship tracks are similar to controlled experiments and allow quantifying aerosol impacts on clouds with exceptional reliability, although experiment-like settings have been created without an intervention by a researcher. The project results will lead to more reliable climate projections urgently needed for climate change mitigation and adaptation.      

Supervisors: Siiri Salupere, Marti Jeltsov, Marko Kaasik

An integral radiation monitoring plan is an important element of any nuclear power program to achieve the fundamental safety objective – protection of people and the environment from the harmful effects of ionizing radiation. Small and modular reactors (SMRs) have become a choice for several utilities in different countries. Yet, integral radiation monitoring programs, which are fit for SMRs, are not currently available for practical use.

This PhD project will develop an integral radiation monitoring plan optimised for SMRs that includes specific instructions for i) pre-operational (construction, background measurement), ii) operational (normal and emergency conditions), and iii) post-operational (decommissioning) phases – so called “greenfield-to-greenfield" approach.

Special attention will be given to off‑site emergency preparedness and response (EPR) arrangements tailored to the scale and needs of SMRs. Focus is placed on the modelling of air pollution dispersion based on atmospheric dynamics. 

The monitoring plan developed in this project will be based on SMRs parameters and deploy modern measurement systems (drones, measurement grid) in combination with real-time dispersion analysis (JRODOS, SILAM etc) to inform EPR activities (sheltering, evacuation, iodine prophylaxis etc.).

Supervisor: Jörg Pieper

In the recent years, effects of climate change including drought and high-light conditions have become more abundant. Therefore, an adaptation to these changing environmental conditions will become unavoidable for a sustainable development of agriculture in the future. One promising strategy is to study native adaptation mechanisms in photosynthesis to battle environmental stress. In this regard, two promising bioprotectants are trehalose and glycerol, which act as stabilizer and plasticizer, respectively, for the photoactive proteins responsible for photosynthesis in plants. 

Trehalose is a specific sugar found in organisms able to survive extreme external stresses, such as high or very low temperatures or periods of complete drought. Most importantly, trehalose was shown to preserve the function of native photosynthetic proteins upon dehydration. Glycerol is a highly hydrophilic compound that is able to attract water, thus, leading to an increase of the protein surface hydration. This plasticizing effect ensures the functionally important protein flexibility and prevents harmful aggregation or misfolding of proteins. However, the molecular mechanisms of the protective roles of trehalose and glycerol are still debated and often poorly understood.

Neutron scattering methods are well-suited for direct nanoscale investigations of protein structure and mobility under nearly native conditions as well as to study their interactions with bioprotectants. Small angle neutron scattering with contrast variation will be used to investigate the location of the trehalose or glycerol molecules with respect to the protein surface and the hydration shell to shed more light on the particular molecular interaction mechanisms. In addition to structural integrity, proteins also have to preserve a specific level of flexibility to perform their function. Quasielastic neutron scattering spectroscopy will be employed to directly study the protein flexibility under environmental stress conditions in the presence of bioprotectants. 

Supervisors: Vitali Nagirnõi, Marco Kirm

The project is aimed at the development of a novel scintillation material for gamma-radiation detection, operating on a combination of Cherenkov emission and ultra-fast cross-luminescence. Such scintillator may be prospective for applications in time-of-flight positron emission tomography (TOF-PET) devices, bringing medical diagnostics on a new quality level with improved spatial resolution and reduced patient dose. The study will be focused on synthesis and thorough investigation by the methods of time-resolved spectroscopy of ternary wide-gap fluorides, containing cations characteristic of cross-luminescence materials (Rb, Cs, Ba) and heavy elements with atomic number exceeding 70 (Lu, Hf, Ta, W, Pb, Bi) facilitating higher yield of Cherenkov emission. The brightest compounds with suitable luminescence properties will be selected for future applied studies. The work will be conducted in close international collaboration with scientists (CCC collaboartion at CERN) specialized in the TOF-PET research and development.

Supervisors: J.M. Kahk, M. Kirm, J. Lischner

The PhD project "First Principles Electronic Structure Calculations for Determining the Principle of Operation of Multicomponent Ultrafast Scintillators," addresses the fundamental properties of inorganic hexafluoride scintillators (K2GeF6, K2SiF6, BaGeF6). The PhD student will be supervised by Dr. Juhan Matthias Kahk, Professor Marco Kirm, and Professor Johannes Lischner. The doctoral project is part of the Horizon Europe Twinning Project EXANST, a collaboration between the University of Tartu, MAX-IV Laboratory, Forschungszentrum Jülich, and Imperial College London.

The project aims to elucidate the operating principles of the specified scintillators, focusing on scintillation yield and response time. The PhD student will use first-principles electronic structure methods such as Density Functional Theory and Green’s Function Theory (GW, Bethe-Salpeter equation) to calculate theoretical spectra of these materials, including photoemission, optical absorption, and luminescence spectra. The results will be compared to relevant experimental results. 

If the computational methods are able to successfully replicate the experimental luminescence properties of the hexafluoride scintillators, in the future, similar calculations can be used to identify other promising materials with a high transition probability for the desirable luminescence transition, and low probabilities for competing processes such as other radiative or non-radiative decay channels.

The computational infrastructure comprises the X-ray Spectroscopy Laboratory computing server, UT Rocket cluster, and LUMI High-Performance Computing system. The PhD project is closely related to three financed research projects: the Horizon Europe Twinning Project EXANST, the Center of Excellence in Sustainable Green Hydrogen and Energy Technologies, and the Horizon Europe Staff Exchanges project BETTERXPS.

Supervisors: Miia Rannikmäe, Regina Soobard, Karin Täht

This doctoral thesis investigates the cognitive and affective factors influencing the learning of natural sciences among Estonian high school students. Specifically, the thesis focuses on negative emotions such as anxiety (fear of an impending problem,), which can hinder the expression of a student's actual knowledge and skills in the context of learning, leading to lower performance and, consequently, increasing anxiety. In the context of learning, various types of anxiety have been studied, such as test anxiety, performance anxiety, and science anxiety. On the other hand, the thesis also pays attention to various positive factors such as scientific literacy, self-efficacy, and epistemic beliefs, which, when supported and cultivated, can reduce anxiety. The topic of the doctoral thesis is relevant, considering that the latest PISA results showed a significant decline in Estonian students' performance in natural sciences. Additionally, according to an international study on children's well-being, the well-being of Estonian children as students has also decreased. The thesis examines various aspects of negative emotions, primarily anxiety, in a subject-specific manner in biology, physics, geography, and chemistry. Additionally, it analyzes the results of e-tests related to anxiety and develops a theoretical framework along with practical guidelines for reducing anxiety associated with learning natural sciences. The results of this work, along with practical guidelines for anxiety reduction, are necessary both for shaping educational policy and for practical teaching purposes.

Speciality admission requirements in Science Education :

• the Doctoral project
• admission interview

In the first stage of assessment, the doctoral thesis project will be assessed. In the second stage of assessment, there is an interview.

Both the doctoral thesis project and admission interview are assessed on a scale of 0 to 50 points, the minimum positive score is 35 points. Each components gives 50% of your overall score with a maximum overall score of 100 points. If you score at least 35 points out of 50 for the thesis proposal, you will be invited to attend an entrance interview.

Requirements for the content and form of the doctoral thesis. Here you will also find who to contact with questions.

The lenght of the doctoral thesis project is up to five A4 pages (without bibliography).

The project presents

·         the research challenge,

·         explains its innovativeness and importance, based on the scientific literature,

·         the research methodology proposed and the resources needed to address it,

·         describes how to expected results are relevant in practice and on the international scientific community.

The candidate is provided with information on the basis of which the candidate can get an overview of the candidate's research potential:

·         information on previous scientific activities (incl. working in positions containing research tasks),

·         a list of published scientific publications and presentations at conferences.

Doctoral thesis project assessment criteria (at least 35 points are required to fulfill the acceptance condition):

1. innovativeness and importance of the doctoral thesis (up to 10 p);

2. justification of the research problem (up to 10 p);

3. adequacy of research methodology (up to 10 p);

4. feasibility of the design (up to 5 p);

5. candidate's research potential for a design-based doctoral thesis (up to 15 credits).

During the admission interview, the candidate will interpret his/her doctoral thesis in English for 5 minutes, followed by an interview with the committee for up to 25 minutes.

Criteria for the evaluation the interview (at least 35 points are required to meet the admission requirement):

1. ability to dechipher the plan of the doctoral thesis (incl. research problem, related scientific literature, methodology) (up to 15 p);

2. motivation to study for a doctorate in education and to work in the field (up to 5 credits);

3. wider analytical and generalization skills in pedagogical topics (up to 10 p);

4. orientation in the problems of Estonian and world educational life (up to 10 p);

5. self-expression skills (all candidates have the ability to express themselves in English, candidates who speak Estonian as their mother tongue also have the ability to express themselves in Estonian) (up to 10 p).

Speciality representative: Miia Rannikmäe, miia.rannikmae@ut.ee

Supervisors: Heili Kasuk, Jack Holbrook

The focus of the research project is the investigation of the impact of science shows on student science learning. By investigating the effectiveness of engaging presentation techniques and memorable demonstrations, the study uncovers short-term and long-term effects on students' scientific literacy, cognitive learning, and career awareness. Additionally, it seeks to determine how phenomena-based science shows foster critical thinking for responsible decision-making. The research methodology involves administering validated instruments to students aged 14-16, complemented by small group interviews. Ultimately, this study aims to enhance understanding of the educational benefits derived from science shows.

Speciality admission requirements in Science Education :

• the Doctoral project
• admission interview

In the first stage of assessment, the doctoral thesis project will be assessed. In the second stage of assessment, there is an interview.

Both the doctoral thesis project and admission interview are assessed on a scale of 0 to 50 points, the minimum positive score is 35 points. Each components gives 50% of your overall score with a maximum overall score of 100 points. If you score at least 35 points out of 50 for the thesis proposal, you will be invited to attend an entrance interview.

Requirements for the content and form of the doctoral thesis. Here you will also find who to contact with questions.

The lenght of the doctoral thesis project is up to five A4 pages (without bibliography).

The project presents

·         the research challenge,

·         explains its innovativeness and importance, based on the scientific literature,

·         the research methodology proposed and the resources needed to address it,

·         describes how to expected results are relevant in practice and on the international scientific community.

The candidate is provided with information on the basis of which the candidate can get an overview of the candidate's research potential:

·         information on previous scientific activities (incl. working in positions containing research tasks),

·         a list of published scientific publications and presentations at conferences.

Doctoral thesis project assessment criteria (at least 35 points are required to fulfill the acceptance condition):

1. innovativeness and importance of the doctoral thesis (up to 10 p);

2. justification of the research problem (up to 10 p);

3. adequacy of research methodology (up to 10 p);

4. feasibility of the design (up to 5 p);

5. candidate's research potential for a design-based doctoral thesis (up to 15 credits).

During the admission interview, the candidate will interpret his/her doctoral thesis in English for 5 minutes, followed by an interview with the committee for up to 25 minutes.

Criteria for the evaluation the interview (at least 35 points are required to meet the admission requirement):

1. ability to dechipher the plan of the doctoral thesis (incl. research problem, related scientific literature, methodology) (up to 15 p);

2. motivation to study for a doctorate in education and to work in the field (up to 5 credits);

3. wider analytical and generalization skills in pedagogical topics (up to 10 p);

4. orientation in the problems of Estonian and world educational life (up to 10 p);

5. self-expression skills (all candidates have the ability to express themselves in English, candidates who speak Estonian as their mother tongue also have the ability to express themselves in Estonian) (up to 10 p).

Speciality representative: Miia Rannikmäe, miia.rannikmae@ut.ee

Supervisors: Lea Hallik, Margit Aun

Healthy wetlands provide numerous ecosystem services such as water purification, flood control, and carbon sequestration. Wetlands play a critical role in mitigating climate change impacts like floods and droughts. Understanding how restored wetlands respond to climate change can inform adaptation strategies for both wetlands and surrounding communities. Earth Observation (EO) data from satellites can cover vast areas quickly and repeatedly, offering a cost-effective way to monitor restoration progress across entire landscapes in a standardized way facilitating comparisons and knowledge sharing. The project will contribute to the development of standardized EO-based protocols for wetland monitoring, which can be easily adopted by other restoration projects and agencies. COPERNICUS Services provide unprecedented amount of temporally and spatially continuous data. Climate predictions are provided at various timescales from seasonal forecasts to long-term climate projections for different scenarios. Re-analysed databases such as ERA5 provide spatially continuous time-series of historic climate data. Earth Observation satellites provide long time-series of monitoring data and derived products (e.g. land cover classes, biophysical products, phenology metrics). Combining EO data with past climate information can help to identify key drivers influencing restoration success, like precipitation patterns, temperature changes, or human activities. Integrating climate projections with EO data can predict potential challenges for the restored ecosystem, allowing for proactive adaptation strategies. The combined analysis of EO and climate data can deepen our understanding of how wetlands respond to restoration efforts and climate change, providing valuable insights for future strategies of climate change adaptation.

Supervisor: Mihkel Pajusalu

The goal of this PhD project is to develop a digital representation of a planetary rover in order to minimise the risk of failure of the system on the Lunar surface and to use it as a monitoring tool throughout its life cycle. In addition, investigation of the use of the digital twin as a tool to optimise and design the physical and software related aspects of the rover will be conducted. The major part of the project will be devoted to improving and developing the Tartu Observatory ULYSSES Moon simulator toolkit to be used as the environment for the digital counterpart. Later in the project a physical counterpart will be added to verify the digital twin accuracy. This project allows the potential PhD student to work in many different theoretical and practical problem areas like graphical user environment development, robotic system interfaces and continuous data acquisition.

Supervisors: Indrek Vurm, Anna Aret

Superluminous supernovae are among the most energetic explosions in the Universe, making them detectable to great distances and hence potentially relevant to various fields of astronomical and cosmological research. While both the superluminous as well as the more ordinary core-collapse supernovae arise from the explosion of a massive star at the end of its life, there is strong evidence to suggest that their energy sources are fundamentally different. A leading candidate for powering the superluminous events is the compact, rapidly spinning and highly magnetized remnant of the stellar core collapse – a magnetar, which continues to deposit energy into the exploding stellar material for weeks to months after the initial explosion. The proposed PhD project focuses on improving our understanding of the physical and radiative mechanisms relevant to shaping the observable radiation of superluminous supernovae in the context of the magnetar model. To this end, we will use detailed numerical simulations following the methods developed at Tartu Observatory to establish a physically motivated link between the central energy source hidden behind several solar masses of progenitor material and the multiwavelength radiation that eventually emerges from the ejecta. A systematic exploration of superluminous supernovae with a detailed physical model will provide insights into their explosion mechanisms, progenitor properties and environments, and will contribute towards utilizing them as new cosmic beacons for probing the distant Universe.

Supervisor: Jan Pisek

Leaf angle distribution (LAD) is a key parameter in models useful for understanding vegetation canopy processes of photosynthesis, evapotranspiration, radiation transmission, and spectral reflectance. Yet, despite the strong sensitivity of many models to variability in LAD, the difficulty in measuring LAD causes it to be one of the most poorly constrained parameters. Satellite remote sensing is currently the only technology able to provide consistent data over large areas and longer periods of time. Multi-angle remote sensing enables us now to describe surface properties by means that are not possible using mono-angle data. This doctoral thesis will devise, test, and implement a novel methodology exploiting multi-angle sensors to retrieve LAD. Improving information about LAD is essential for advancing ecological understanding of its role within the biophysical interaction of sunlight and the vegetation canopy, and would improve prediction and forecast horizons of vegetation dynamics globally.

Supervisors: Elmo Tempel, Taavi Tuvikene

The doctoral thesis is devoted to studying the cosmic web environment and galaxy properties, using data from the state-of-the-art spectroscopic sky surveys WAVES and 4HS within the 4MOST project. These galaxy surveys will bring a huge increase in the number of galaxies with known distances, revealing the cosmic web with unprecedented details. These details allow us to study the physical and environmental processes behind galaxy evolution via rigorous statistical analysis and comparison with state-of-the-art hydrodynamical galaxy evolution simulations. An integral part of the thesis is developing new analysis methods to overcome observational biases and make full use of the statistical power of the surveys.

Supervisors:  Rutha Jäger, Jaak Nerut, Enn Lust

Hydrogen is essential for various applications, such as a fuel in transport systems and as a feedstock in the chemical industry, and its global demand is projected to double from 94 to 180 million tonnes by 2030. Presently, hydrogen is primarily produced from fossil fuels, resulting in significant CO2 emissions. To address this, electrolysis of water, powered by renewable electricity, offers a sustainable solution. Proton exchange membrane electrolysis (PEMEL) stands out for its efficiency. However, its industrialisation is still in its early stages. PEMEL relies on platinum group metals (PGMs) like Pt and Ir (or Ru) for electrode materials, raising concerns about PGM shortages with increased PEMEL production capacity. Thus, there is a critical need to develop novel electrode materials either with reduced PGM content or introduce non-PGM catalysts. Incorporating rare earth elements into Ir or Ir-Ru oxides is an effective strategy for modifying their electronic structure, thereby enhancing catalytic activity and durability in PEMEL. The project aims to develop and assess innovative catalysts based on rare earth metal oxides for PEMEL use. This involves conducting physical characterisation alongside electrochemical analysis to comprehend their performance in PEMEL systems. The most effective catalysts are applied to complete a PEMEL stack.

Juhendajad: Rasmus Palm, Enn Lust, Jaak Nerut, Angélica María Baena Moncada

Metal-organic frameworks (MOFs) include a wide variety of meso- and microporous structures with the potential to incorporate different metals and heteroatoms. Carbonisation of MOFs yields highly porous carbon materials with uniformly distributed catalytically active functional centres. The aim of this PhD project is to synthesize and carbonise MOFs yielding highly functional catalytically active model carbon materials. The MOF derived materials will be used to investigate the effect of different active sites in a carbon material to the adsorbate/adsorbent interactions important for energy storage and energy conversion characteristics. The focus of this PhD project will be to use such MOF derived carbon materials to investigate the beforementioned effect of active sites on adsorbate/adsorbent interactions crucial for energy storage and conversion applications. This includes the use of MOF derived carbons to investigate the effect of different active sites on hydrogen adsorption and hydrogen spillover effect, on the catalytic hydrogen storage properties of nanoconfined hydrides, and on the catalytic activity of oxygen reduction reaction kinetics at activated materials. The PhD project will involve the use of various experimental methods like gas adsorption, x-ray diffraction, Raman spectroscopy, and more. In addition, neutron scattering methods will be used to investigate the structures and processes of importance in/at the MOF derived carbons for energy storage and conversion applications.