THESIS
2018
xvi, 154 pages : illustrations (chiefly color) ; 30 cm
Abstract
Understanding the structure and dynamics of the biological macromolecules is the central
pillar of molecular biology and biochemistry. Even with the great advances in both
experimental and computational techniques in the previous decades, one still often encounters
cases where the required spatial and temporal resolution could not be attained by existing
tools, which greatly limits our understanding of many biologically relevant processes. In this
work, I have presented several algorithmic developments that could help us bridge the gap
between computation and experiment through a better utilization of simulation or
experimental data. In particular, some of the works are dedicated to the development of
techniques that extends the applicability of Markov State Model, a stochastic...[
Read more ]
Understanding the structure and dynamics of the biological macromolecules is the central
pillar of molecular biology and biochemistry. Even with the great advances in both
experimental and computational techniques in the previous decades, one still often encounters
cases where the required spatial and temporal resolution could not be attained by existing
tools, which greatly limits our understanding of many biologically relevant processes. In this
work, I have presented several algorithmic developments that could help us bridge the gap
between computation and experiment through a better utilization of simulation or
experimental data. In particular, some of the works are dedicated to the development of
techniques that extends the applicability of Markov State Model, a stochastic model that
allows the prediction of long-timescale dynamics based on relatively short MD simulations.
All these efforts would have extended the kinetics prediction to a much longer timescale using
the same MD data. Another important part of the presented work involves a better utilization
of Cryo-EM images obtained experimentally to infer the underlying free energy landscape. A
two-stage classification scheme has been presented that could accurately classify the Cryo-EM images from multiple conformations, such characterization allows the determination of relatively free energy of different conformations simply by counting. The efforts presented in this work are expected to close the gap between experiments and simulations by expanding
the applicability of both regimes.
Post a Comment