Open Collections

Design space exploration in ion channels using fine grained brownian dynamics

University of British Columbia

Design Space Exploration (DSE) in the context of ion channels refers to the systematic exploration of a design space defined with the dimensions of the space corresponding to channel characteristics. The goal of DSE here is to find points within the space that maximize a figure of merit related to conduction. Finding an efficient means to performing DSE for ion channels holds promise in several application areas such as nano-medicine and drug development, where it is commonly desirable to efficiently reverse-engineer drugs or channels in order to determine which channel structures or drugs lead to a particular conduction behavior. One example of DSE would have the dimensions of the design space specified by dielectric constants throughout the channel, and with the figure-of-merit defined by conduction. The primary roadblock to using DSE for ion channels is computational complexity as evaluating each channel characteristic (design point) requires 10¹⁰ simulation iterations. If, for example, the design space is defined by 5 parameters each having 10 possible values, the process of evaluating all possible combinations of these parameters exhaustively would require 10⁵ * 10¹⁰ = 10¹⁵ simulation iterations. Depending on the time it takes for each iteration, a DSE study could take years or even decades. As a result, it is critical that the approach used for evaluating each design point is both fast and efficient in order to save on simulation time and computational resources. This thesis proposes two-fold strategy for improving the efficiency of DSE for ion channels. First, it proposes an approach for improving the speed of DSE by systematically reducing the design space size using statistical-based inference. It shows how this methodology can be utilized to reduce the design space size by orders of magnitude for two different scenarios: with and without the presence of a drug in the channel. Second, it proposes a novel Finite Grained Brownian Dynamics framework for evaluating design points. Using both approaches together, the framework achieves accuracy that is consistent with Molecular Dynamics (with R² = 82%), a significantly higher resolution modeling technique, at a fraction of the cost.

Text

2020-03-02

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/73644

Electrical and Computer Engineering

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/