Protein Structural Topology Comparison

Abstract

In the summer of 2016, I worked with the Computational Biology group as an algorithms developer. Understanding protein structural topology provides an important baseline to studying protein functions and their interactions. Majority of approaches for protein structure build upon their the primary sequence prediction. State of art in primary sequence prediction is a lossy process and hence stochasitically impacts structure understanding and the subsequent study of protein function and protein-protein interaction.
I worked on an algorithm to develop a metric for primary-sequence independent protein structural topology comparison. The approach built upon the representation of protein in vector space, taking into consideration conformational transformations, to dynamically determine pairwise similarity. Visualising the potential impact few hundred lines of dynamic programming code could have left me amazed.

Reference to code: [Github]