Comparative Markov state analysis of APOE protein dynamics by neural networks
| Název práce v češtině: | Srovnávací analýza markovských stavových modelů pro dynamiku proteinu APOE pomocí neuronových sítí |
|---|---|
| Název v anglickém jazyce: | Comparative Markov state analysis of APOE protein dynamics by neural networks |
| Klíčová slova: | Strojové učení pro molekulární dynamiku|Neuronové sítě|Variační přístup k Markovským procesům|Modely Markovských stavů|APOE protein |
| Klíčová slova anglicky: | Machine learning for molecular dynamics|Neural networks|Variational approach to Markov processes|Markov state models|APOE protein |
| Akademický rok vypsání: | 2022/2023 |
| Typ práce: | diplomová práce |
| Jazyk práce: | angličtina |
| Ústav: | Katedra softwaru a výuky informatiky (32-KSVI) |
| Vedoucí / školitel: | Mgr. Jiří Sedlář, Ph.D. |
| Řešitel: | skrytý - zadáno a potvrzeno stud. odd. |
| Datum přihlášení: | 28.03.2023 |
| Datum zadání: | 30.03.2023 |
| Datum potvrzení stud. oddělením: | 12.07.2023 |
| Datum a čas obhajoby: | 05.09.2023 09:00 |
| Datum odevzdání elektronické podoby: | 20.07.2023 |
| Datum odevzdání tištěné podoby: | 24.07.2023 |
| Datum proběhlé obhajoby: | 05.09.2023 |
| Oponenti: | prof. RNDr. Ing. Martin Holeňa, CSc. |
| Konzultanti: | Josef Šivic |
| Zásady pro vypracování |
| The objective of this master's thesis is to investigate the molecular dynamics of Apolipoprotein E (APOE) by neural networks. APOE is a protein associated with the development of Alzheimer's disease. Insight into the APOE dynamics is highly relevant for understanding the molecular basis of this neurodegenerative disease and estimating the effect of drug candidates.
The variational approach for Markov processes [1] by neural networks (VAMPnets) has proved useful in analyzing the dynamics of small proteins [2, 3, 4]. For example, the Comparative Markov State Analysis approach (CoVAMPnet) [5] was used to evaluate the effect of drug candidates on the dynamics of amyloid-beta42 (Abeta42), a small disordered protein (42 residues) related to the onset of Alzheimer's disease. While the existing VAMPnet-based architectures have been used primarily for such small proteins, the APOE protein is much larger (ca. 300 residues). This thesis aims to analyze the APOE protein dynamics and the effect of a drug candidate (small molecule) on the dynamics. The project's objectives are to: 1. Read and understand the papers on the variational approach for Markov processes [1] and VAMPnet-based neural networks [2-5]. 2. Analyze the dynamics of the APOE protein using the CoVAMPnet approach [5]. 3. Analyze the effect of a small molecule drug candidate on the dynamics of the protein using the CoVAMPnet approach [5]. 4. Optionally explore the division of the APOE protein into dynamically independent domains. This could result in a more interpretable model of its dynamics and reduced time complexity. Potentially explore also iVAMPnet [6], a promising new architecture that identifies such domains automatically. Note that it is unclear at this point how dynamically independent the APOE domains are. The thesis will be co-advised in collaboration [5] between the Czech Institute of Informatics, Robotics and Cybernetics at the Czech Technical University in Prague providing expertise in machine learning and the Loschmidt Laboratories at Masaryk University in Brno providing expertise in protein engineering. |
| Seznam odborné literatury |
| Literature:
[1] H. Wu, F. Noé. Variational Approach for Learning Markov Processes from Time Series Data. J Nonlinear Sci 30, 23-66 (2020). https://doi.org/10.1007/s00332-019-09567-y [2] A. Mardt, L. Pasquali, H. Wu, F. Noé. VAMPnets for deep learning of molecular kinetics. Nat Commun 9, 5 (2018). https://doi.org/10.1038/s41467-017-02388-1 [3] T. Löhr, K. Kohlhoff, G.T. Heller, C. Camilloni, M. Vendruscolo. A kinetic ensemble of the Alzheimer’s Aβ peptide. Nat Comput Sci 1, 71–78 (2021). https://doi.org/10.1038/s43588-020-00003-w [4] T. Löhr, K. Kohlhoff K, G.T. Heller, C. Camilloni, M. Vendruscolo. A Small Molecule Stabilizes the Disordered Native State of the Alzheimer's Aβ Peptide. ACS Chem Neurosci 13, 12, 1738-1745 (2022). https://doi.org/10.1021/acschemneuro.2c00116 [5] S.M. Marques, P. Kouba, A. Legrand, J. Sedlar, L. Disson, J. Planas-Iglesias, Z. Sanusi, A. Kunka, J. Damborsky, T. Pajdla, Z. Prokop, S. Mazurenko, J. Sivic, D. Bednar. Effects of Alzheimer’s Disease Drug Candidates on Disordered Aβ42 Dissected by Comparative Markov State Analysis CoVAMPnet. bioRxiv preprint (2023). https://doi.org/10.1101/2023.01.06.523007 [6] A. Mardt, T. Hempel, C. Clementi, F. Noé. Deep learning to decompose macromolecules into independent Markovian domains. Nat Commun 13, 7101 (2022). https://doi.org/10.1038/s41467-022-34603-z |