Single-cell analysis in systems immunology – an application of novel unsupervised tools in infectious diseases and cancer

• Název práce v češtině: Jednobuněčná analýza v systémové imunologii - aplikace nových nesupervizovaných nástrojů v infekčních a nádorových onemocněních
• Název v anglickém jazyce: Single-cell analysis in systems immunology – an application of novel unsupervised tools in infectious diseases and cancer
• Klíčová slova: imunitní odpověď, nesupervizovaná analýza, redukce dimenzionality, embedding, hierarchické klastrování, topologická analýza dat, strojové učení, klinická data, infekce, nádor
• Klíčová slova anglicky: immune response, unsupervised analysis, dimensionality reduction, embedding, hierarchical clustering, topological data analysis, machine learning, clinical data, infectious disease, tumor
• Akademický rok vypsání: 2021/2022
• Typ práce: disertační práce
• Jazyk práce: angličtina
• Ústav: Katedra buněčné biologie (31-151)
• Vedoucí / školitel: RNDr. Karel Drbal, Ph.D.

Zásady pro vypracování

MB151P108 Clinical Cases in Immunology / Jakub Abramson
MB151P103E Immunology - a systems biology view / Karel Drbal
MB151P107 Protein dynamics in development and cancer / Lukáš Čermák
MB151P80E Cytometry / Karel Drbal
MB140P86 Methods of functional genomics / Vladimír Beneš / Martin Pospíšek
MB151P112E Genome-scale metabolic models (EMBL Heidelberg)
MB151P113E Analytical methods in cancer and population genomics and transcriptomics (EMBL Heidelberg)
MB151P115E Spatiotemporal modeling and simulation of biological systems (MPI Dresden)
MB151P114E Biological imaging and image analysis (MPI Dresden)

Seznam odborné literatury

Christopher Gandrud: Reproducible Research with R and RStudio, 3rd Edition (Chapman &Hall/CRC)
Inna Kuperstein, Emmanuel Barillot: Computational Systems Biology Approaches in Cancer Research (Chapman & Hall/CRC Computational Biology Series)
Uri Alon: An Introduction to Systems Biology: Design Principles of Biological Circuits, Second Edition (Chapman &Hall/CRC Mathematical and Computational Biology)

Předběžná náplň práce

Budeme aplikovat nové nástroje pro analýzu velkých transkriptomických a proteomických dat v klinice způsobem nesupervizované rychlé topologické analýzy dat (TDA), která je založená na klastrování a vizualizaci statisticky hodnocených dat. Tato metoda poskytuje vysokou rychlost pro redukci dimenzionality až na 2D. Předmětem této práce je optimalizace parametrů a zajištění minimální distorze dat a vysoké reproducibility. Pracujeme v prostředí Galaxy, nicméně detailní znalost R je nutnou podmínkou (programování v C++ je výhodou). To nám poskytuje naprosto nový pohled na vědecká data obecně a umožňuje nám jejich reinterpretaci a objev nových souvislostí.
Soustředíme se na data klinického charakteru od pacientů s infekční nebo nádorovou chorobou. Vlastní sběr dat a jejich databázová organizace je součástí této práce. Dostupnost experimentálních modelů zebřičky a medaky umožňuje následnou validaci predikovaných směrovaných příčinných vztahů na úrovni transkriptomu a proteomu.

Předběžná náplň práce v anglickém jazyce

The objective of this project is an application of novel analytical tools for the large transcriptomic and proteomic datasets in clinics. This is in principle unsupervised topological data analysis (TDA) based on clustering, which allows for immediate statistical data evaluation and visualization using dimensionality reduction down to 2D. We are going to optimize parameters, achieve minimal data distortion and maximal reproducibility as well. Galaxy platform is a central cloud environment, however, a deep knowledge of R/Python is essential (C++ programming is a bonus). In turn, this brings a completely new understanding of existing scientific data in general and allows for the reinterpretation and discovery of new relationships.
We will focus on clinical datasets in patients suffering from infectious or tumor diseases. An inherent part of the workflow is data collection and database maintenance. A prediction of directional causal relationships will be finally validated in available zebrafish and/or medaka models in our laboratory at the level of transcriptome and proteome.
Recently, an excessive boost in the development of analytical bioinformatic algorithms allows biologists to mine the available datasets originating from single cells in a completely unsupervised manner for the first time. The human body is composed of around 30 trillion cells and the objective of this application is the use of novel dimensionality reduction, trajectory inference, and clustering algorithms in order to decode their directional relationship in the field of systems immunology.
Our research focuses on dynamic systems of immune response monitoring in patients suffering from infectious diseases – tuberculosis or borreliosis – as well as various solid tumors – mainly bladder cancer. Under these pathological conditions, activation of immune cell subsets of both innate and adaptive systems regulate the outcome of the disease. A single-cell oriented statistical data evaluation and visualization finally stratify each patient. We are going to optimize the parameters of the non-linear computational methods in order to preserve data distribution and maximize reproducibility.
An inherent part of the workflow is the collection of genomic/transcriptomic/proteomic data and patient database maintenance. The predictions of directional causal relationships will be validated in available zebrafish and/or medaka models after multiparametric flow cytometry and cell sorting in our laboratory at the level of gene and cellular networks. A collaboration with local clinical partners (TH, HNB, Prague) and computation centers (IDA FEE CTU [1], IOCB CAS [2], Prague) are backed by recent publications. International collaboration within LifeTime consortia will be an inherent part of the project.
As stated above, deep knowledge of R and Python (C++ is a bonus) and immunology is an essential profile of a successful candidate. Optionally, the experience with cytometry and/or microscopy and the knowledge of one of the experimental models is a plus. Two major goals of this Ph.D. position are 1/ the integration of existing tools into a Galaxy pipeline or development of a standalone application and 2/ the identification of cellular biomarkers of either latent TB infection, late-stage Lyme disease or bladder cancer stem cells.
1. Dvorakova, E. et al. Bioinforma. Res. Appl. ICBRA 2019 (2019). at http://ida.felk.cvut.cz/zelezny/pubs/icbra2019.pdf
2. Kratochvíl, M. et al. bioRxiv 496869 (2019).doi:10.1101/496869