Modeling cellular metabolism in multiple scales

Computational studies of biological systems have evolved significantly, to the point where we can combine quantitative information on the expression of thousands of proteins and RNAs in large models of metabolism or gene expression. We can recover and explain fluctuations observed in single-cell experiments, and connect them to physiological behavior. In order to advance the field, it is necessary to employ state-of-the-art computational tools, and integrate even more complex biological contexts into the models and modeling tools themselves. In order to enhance our description of physiological variability in metabolic phenotypes of yeast, we improved the Population FBA method to dynamically constrain a genome-wide metabolic reconstruction of S. cerevisiae using comprehensive proteomics and microarray data sets. Moreover, a genetic algorithm approach was used to explore the variability of observed metabolic phenotypes, identifying a surprising robustness in overall growth rate distribution. Within cell-wide metabolic or gene expression networks, key regulatory checkpoints are modulated by specific molecular interactions. They are essential for the formation of multi-protein complexes, for the binding of ligands in active sites, and for allosteric activation of enzymes. In order to identify essential residues for communication pathways and signaling elements in biomolecular complexes, we developed an evolution of the dynamical network analysis framework. The new implementation was applied to the leucine tRNA synthetase complexed with its cognate tRNA and adenylate, revealing experimentally verified identity elements throughout the tRNA:protein interface, as well as essential catalytic residues in the active site. This enhanced and updated methodology will provide the community with an intuitive and interactive interface, that can be easily applied to large macroolecular complexes. Beyond the study of protein:ligand interactions, in order to analyze specific enzymatic reactions, one needs to combine quantum mechanics (QM) and molecular mechanics (MM) methods to secure sub-atomic resolution within relevant biological time and length scales. Despite recent advances QM calculations, the computational cost of studying nanosecond-long dynamics of entire systems relying solely on QM methodologies is prohibitive. To circumvent this cost barrier, it is possible to confine the QM formalism to a subregion of a system and to include the effects of the surrounding system through MM simulations, leading to hybrid “QM/MM” simulations. Unfortunately, a majority of the available QM/MM implementations lack a comprehensive set of features that could make these calculations even more attractive. In this work, we chose NAMD to be the center of a comprehensive QM/MM suite that provides easy setup, visualization and analysis of QM/MM simulations, while supporting the simulation of many independent QM regions, and a smooth integration with a collection of state-of-the-art free energy and enhanced sampling methods.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Biophysics & Computnl Biology

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2019

Author dc:creator

• Cardoso Dos Reis Melo, Marcelo

Contributors dc:contributor

• Luthey-Schulten, Zaida
• Martinis, Susan
• Tajkhorshid, Emad
• Dar, Roy

• system's biology
• qm/mm
• protein:trna interaction

Statement dc:rights

• Copyright 2018 Marcelo Cardoso dos Reis Melo

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/105124

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/105124