University of Texas Southwestern Medical Center
Complexity is a key obstacle in understanding and engineering biological systems. The details of this complexity rapidly accumulate due to advances in sequencing, imaging, and multiplexing. This leads to larger and more precise datasets. Longstanding and widely-applicable computational approaches, such as mutual information analysis, principal components analysis, and machine learning are standard tools currently used for data analysis. However, these standard tools are limited in matching a methodology to a data structure or biological process and typically do not lead to a mechanistic or causative understanding. Therefore, it is increasingly important to apply problem-specific yet unbiased methods to extract robust features from biological data and to fit the data within a predictive conceptual framework. This collection brings a wide range of computational and theoretical methods to extract underlying patterns and organizational principles in molecular biology that would otherwise be obscured by complexity. While these methods draw on insights from multiple disciplines, they all aim to simplify complex phenomena by finding sparse and modular representations that take into account the constraints of molecular biology.