|
|
The Application of Variational Autoencoders for Ancestral Sequence Reconstruction
Kohout, Pavel ; Martínek, Tomáš (referee) ; Musil, Miloš (advisor)
Protein engineering is an interdisciplinary science concerned with the design of improved proteins. A successful method used to design more stable and active proteins is ancestral sequence reconstruction. This method explores the evolutionary relationships between existing proteins and uses phylogenetic trees to generate their evolutionary ancestors, which often exhibit the desired improved properties. Therefore, new and more robust methods using mathematical models together with huge amounts of sequence data could become a powerful tool for protein engineering. This thesis explores the use of variational autoencoders as an alternative approach to ancestral sequence design compared to conventional methods using phylogenetic trees. Experiments were performed to optimize the architecture and statistical methods were proposed to evaluate the quality of the models and the sequences generated. At the same time, robustness tests of the whole method were performed and strategies for ancestral sequence generation were proposed and implemented.
|
|
|
Machine Learning as a Tool for the Prediction of the Effect of Mutations on Protein Stability
Dúbrava, Juraj Ondrej ; Martínek, Tomáš (referee) ; Musil, Miloš (advisor)
The main focus of this thesis is the prediction of the effect of amino acid substitutions on protein stability. My goal was to develop a predictive tool for the classification of the effect of mutations by combining several machine learning techniques. The implemented predictor, which utilizes SVM and Random forest methods, has achieved higher accuracy than any of the integrated methods. The novel predictive tool was compared with the existing ones using independent testing dataset. The predictor has yield 67 % accuracy and MCC 0,3.
|
|
|
Large-Scale Analysis of the Ligand Transport and Docking inside of the Protein Tunels
Ježík, Andrej ; Martínek, Tomáš (referee) ; Musil, Miloš (advisor)
This thesis discusses large-scale analysis of the ligand transport and docking inside of the protein tunnels. Protein-ligand interactions are involved in processes such as cell signalling, transport, metabolism, regulation, gene expression, and enzyme activity. To understand the interaction between these molecules is vitally important for the research for new pharmaceuticals. The procedure of protein-ligand docking involves the following steps: (i) finding the structures of proteins (receptors) and ligands, (ii) identifying ligand binding sites, (iii) considering receptor/ligand flexibility, and (iv) computing interaction energy between the receptor and the ligand. Additional functionality will be implemented to allow CaverWeb to test a complete set of pre-processed drug ligands on a protein, in an effort to enhance the efficiency of the procedure for large sets of ligands, which will allow a much smoother workflow.
|
|
|
Interactive Database for the Storage and Maintenance of the Biological Data
Dúbrava, Juraj Ondrej ; Martínek, Tomáš (referee) ; Musil, Miloš (advisor)
Cieľom tejto práce je vytvorenie novej databázy dát pre proteínovú stabilitu, ktorá bude udržiavať a poskytovať experimentálne dáta. Výsledkom práce je databáza FireProtDB, ktorá poskytuje manuálne overené experimentálne dáta z dostupných zdrojov a implementuje grafické užívateľské rozhranie, ktoré poskytuje dôležité informácie o dátach spoločne s možnosťou vyhľadávania umožňujúcim vytvárať dotazy na mieru a cieliacim na užívateľov, ktorí hľadajú dáta pre vytváranie dátových sád pre nástroje využívajúce strojové učenie.
|
|
|
Analysis of Cancer-Associated Protein Mutations
Borko, Simeon ; Martínek, Tomáš (referee) ; Musil, Miloš (advisor)
PredictSNP ONCO is a web tool that analyses the impact of protein mutations on the development of oncological diseases. The tool integrates several bioinformatics tools and provides the result of its own predictor, which classifies mutations as oncogenic or benign. Apart from that, PredictSNP ONCO provides the results of a virtual screening of drugs that could inhibit the function of affected proteins. The web interface is freely available at https://loschmidt.chemi.muni.cz/predictsnp-onco/.
|
|
|
The development of energetics in the Czech Republic during the following years
Musil, Miloš ; Kracík, Petr (referee) ; Toman, Filip (advisor)
This thesis analyses the current Czech energy sector. It describes its current state and outlines possible scenarios of its development in future years in the context of the energy transformation. The thesis seeks to analyse the current situation of the energy sector in the country, assess its strengths and weaknesses and identify opportunities and threats for future development. The thesis describes the different parts of the Czech energy sector, such as the electricity, gas, heating and oil industries, including production sources, consumption and emissions. Most emphasis is given to the electricity sector. The capacities and potentials of the different sectors are analysed alongside the possible future scenarios outlined. Part of the thesis is also focused on describing and explaining the composition and construction of electricity and gas prices in the Czech Republic. Emphasis is also placed on placing the Czech energy sector in the broader context of the energy sectors of European countries and includes various scenarios for the development of the electricity sector in Europe. In the context of the current geopolitical situation in Europe and the modernisation and transformation of this sector, it is necessary to build new energy sources.
|
|
|
Fully Automated Method for the Design of Ancestral Proteins
Štěpánek, Martin ; Martínek, Tomáš (referee) ; Musil, Miloš (advisor)
Stabilita proteinů je zásadní vlastností ovlivňující možnosti průmyslového použití proteinů. Tato práce se zaměřuje na vylepšení automatizovaného nástroje FireProtASR, který využívá rekonstrukci ancestrálních sekvencí pro návrh proteinů se zvýšenou stabilitou. Do tohoto nástroje byla přidána nově navržená metoda rekonstrukce successorů. Successory představují sekvence proteinů, ke kterým bude v budoucnosti směřovat evoluce. Podle současných poznatků by měly mít vyšší aktivitu než současné proteiny, společně s vyšší schopností vázat se specificky na určité molekuly. Dále bylo vytvořeno nové webové uživatelské rozhraní, které poskytuje rychlé a intuitivní ovládání nástroje FireProtASR.
|
|
|
Computational Workflow for the Prediction and Design of Stable Proteins
Kadleček, Josef ; Martínek, Tomáš (referee) ; Musil, Miloš (advisor)
Tato práce se soustředí na rozšíření výpočetního nástroje pro predikci mutací proteinů za účelem zvýšení jejich stability. Stabilizace proteinů je nezbytnou součástí návrhu léčiv, a jelikož experimentální vyhodnocení přínosu jednotlivých mutací na stabilitu proteinu je nákladné a zdlouhavé, byl vyvinut nástroj FireProt. FireProt využívá kombinaci evolučních a energetických přístupů pro identifikaci potencionálně stabilizujících mutací. Tato práce rozšiřuje nástroj FireProt o možnost zadat vlastní mutace k analýze, integruje predikční nástroj FireProt ASR a dává možnost uživateli vybrat z několika strategií pro detekci stabilizujících mutací. Tato práce dále rozšiřuje nástroj FireProt o další informace k jednotlivým residuím (jako je například relativní B-faktor), umožňuje využít homologní modelování k vytvoření struktury proteinu na základě jeho sekvence, a v neposlední řadě představuje nový přístup k návrhu vícebodových mutantů.
|
|
|
The Application of Variational Autoencoders for Ancestral Sequence Reconstruction
Kohout, Pavel ; Martínek, Tomáš (referee) ; Musil, Miloš (advisor)
Protein engineering is an interdisciplinary science concerned with the design of improved proteins. A successful method used to design more stable and active proteins is ancestral sequence reconstruction. This method explores the evolutionary relationships between existing proteins and uses phylogenetic trees to generate their evolutionary ancestors, which often exhibit the desired improved properties. Therefore, new and more robust methods using mathematical models together with huge amounts of sequence data could become a powerful tool for protein engineering. This thesis explores the use of variational autoencoders as an alternative approach to ancestral sequence design compared to conventional methods using phylogenetic trees. Experiments were performed to optimize the architecture and statistical methods were proposed to evaluate the quality of the models and the sequences generated. At the same time, robustness tests of the whole method were performed and strategies for ancestral sequence generation were proposed and implemented.
|
|
|
Identification of the Protein Tunnels Using Molecular Dynamics
Kohout, Petr ; Martínek, Tomáš (referee) ; Musil, Miloš (advisor)
This thesis focuses on the analysis of protein structures. The aim is to design Caver Web 2.0 -- a new version of the web application that integrates additional scientific tools and allows users to go through a complicated workflow to provide relevant results without the need for a deeper knowledge of the integrated tools. Everything will be delivered through a simple and interactive user interface. The application extends the original Caver Web 1.0 application with new features. Caver Web 1.0 is a web server suitable for identifying protein tunnels and channels for which it allows to run ligand transport analyses. The program is characterized by an intuitive and user-friendly interface with minimum required input from the user. The server is suitable for researchers without advanced bioinformatics or technical knowledge. Its current version is well established and highly used in the scientific community (35,000 completed calculations in two years of operation). The most significant limitation of the current version is the ability to analyze only static structure, which often provides an incomplete biological picture. Therefore, it was decided to extend the tool to calculate molecular dynamics to provide a comprehensive picture of protein structure changes.
|
guest ::
RSS feed.