|
Programme Details
| Study Language
|
Czech |
|
Standard study length
|
4 years |
| Form of study
|
combined
,
full-time
|
| Guarantor
|
prof. Mgr. Daniel Svozil, Ph.D.
|
| Place of study
|
Praha |
| Capacity
|
8 students |
| Programme code (national)
|
P0588D030009 |
| Programme Code (internal)
|
D107
|
| Number of Ph.D. topics
|
4 |
Ph.D. topics for study year 2026/27
Bioinformatic analysis of deoxyribozymes
|
Study place:
|
Institute of Organic Chemistry and Biochemistry of the CAS
|
| Guaranteeing Departments: |
Institute of Organic Chemistry and Biochemistry of the CAS
|
|
Also available in study programmes:
|
Bioinformatics (
in English language
)
|
| Supervisor: |
Edward A. Curtis, Ph.D.
|
| Expected Form of Study: |
Full-time |
| Expected Method of Funding: |
Scholarship + salary |
Annotation
Deoxyribozymes are DNA molecules that catalyze reactions. They are useful tools due to their low cost, high stability, and ability to function over a wide range of conditions. The goal of this thesis will be to develop new methods to analyze large deoxyribozyme datasets generated in our lab using selection and high-throughput sequencing. The project will require knowledge of standard bioinformatic approaches as well as emerging methods such as machine learning.
Nuisance Compounds in High-Throughput Screening
|
Study place:
|
Institute of Macromolecular Chemistry of the CAS
|
| Guaranteeing Departments: |
Institute of Macromolecular Chemistry of the CAS
|
|
Also available in study programmes:
|
Bioinformatics (
in English language
)
|
| Supervisor: |
Ing. Ctibor Škuta, Ph.D.
|
| Expected Form of Study: |
Full-time |
| Expected Method of Funding: |
Scholarship |
Annotation
High-Throughput Screening (HTS) is a key technology in the early stages of drug discovery. Interpreting results can be complicated by the presence of nuisance compounds behavior (PAINS, aggregators, autofluorescence, cytotoxicity, and others) that obscure the distinction between true and false positives. Metadata-rich datasets from the CZ-OPENSCREEN and EU-OPENSCREEN infrastructures, as well as large public repositories, such as PubChem, provide opportunity to better understand existing patterns, as well as identify new ones. This context-aware methodology aims to investigate how compound structures, experimental setup, and chosen biological targets relate to observed nuisance behavior. The work involves compiling, annotating, and developing tools for detecting and interpreting nuisance compounds, including substructure filters, curated compound lists, machine-learning models, and their composite approaches. Particular focus is given to distinguishing different nuisance mechanisms, highlighting the likelihood and severity of various flags, and their combinations, and compiling these insights into interpretable, human-readable formats. This work also involves exploring the use of large language models to generate clear, context-sensitive labels and descriptions. These tools can be packaged into open-access resources for compound triage and integrated into the CZ-/EU-OPENSCREEN ecosystem, with the CZ-OPENSCREEN facility offering their experimental validation.
Research and development of protein language models for the purpose of optimization and design of new biologically active compounds
|
Study place:
|
Department of Informatics and Chemistry, FCT, VŠCHT Praha
|
| Guaranteeing Departments: |
Department of Informatics and Chemistry
|
| Supervisor: |
Ing. Martin Šícho, Ph.D.
|
| Expected Form of Study: |
Full-time |
| Expected Method of Funding: |
Scholarship + salary |
Annotation
Biologically active molecules play a crucial role in drug discovery and chemical biology, serving as the foundation for the development of new therapeutics and vaccines. These molecules, such as antibodies, are designed to interact with specific protein targets to modulate biological pathways and treat diseases, making their discovery and optimization a central goal in modern pharmaceutical research. Antibody drugs and vaccines depend on how antibodies interact with their target proteins. Recently, protein language models have emerged as a novel computational method for the advancement of therapeutic antibody discovery, as they can efficiently predict critical epitope characteristics and binding relationships. Utilizing protein langugage models, this dissertation addresses the design and developability of antibodies with special focus on ensuring that the models support practical therapeutic applications. In terms of developability, the project will mainly focus on the prediction of hydrophobicity, self-interaction and thermostability, which are important properties to consider in pre-clinical phases of development. In addition, generative artificial intelligence (AI) models for target-specific design of antibodies were developed. Finally, the dissertation also explores the concept of anomaly detection as a route towards the discovery of unusual and potentially useful candidates for antibody design.
Deciphering Lipid Metabolism in Cancer: Integrative Approaches in Metabolomics, Fluxomics, and Metabolic Engineering
|
Study place:
|
Institute of Physiology of the CAS
|
| Guaranteeing Departments: |
Institute of Physiology of the CAS
|
|
Also available in study programmes:
|
Bioinformatics (
in English language
)
|
| Supervisor: |
RNDr. Ondřej Kuda, Ph.D.
|
| Expected Form of Study: |
Full-time |
| Expected Method of Funding: |
Scholarship + salary |
Annotation
This PhD project investigates the rewiring of lipid metabolic pathways in cancer using an integrative approach combining metabolomics, fluxomics, metabolic engineering, and in silico modeling. The research aims to deconvolute complex lipid metabolic pathways through metabolic flux analysis and stable isotope tracer studies, supported by Python-based data processing pipelines and advanced computational modeling.
The study incorporates experimental work, including cancer cell culture systems and in vivo mouse models, to validate findings and quantify metabolic fluxes under physiological and pathological conditions. In silico simulations of lipid metabolism will be used to predict pathway behavior and identify potential intervention points. Machine learning approaches will aid in biomarker discovery and the prediction of metabolic vulnerabilities, offering insights into the mechanisms driving cancer progression and potential therapeutic targets.
This interdisciplinary project bridges computational biology, biochemistry, and experimental cancer research, contributing to our understanding of lipid metabolism and the development of precision strategies for metabolic engineering and cancer therapy.
The work will be conducted at the IPHYS CAS. The work is financially secured in terms of material and full time position.
|
