Persistent homology for protein folding analysis

Abstract

Persistent homology is a concept in topological data analysis used to reduce the dimensionality and<br>complexity of the data sets, and also to determine topological features and delete noises. In this talk,<br>unfolding process is simulated with the constant velocity pulling algorithm of SMD. Then, we apply<br>the persistent homology to reveal the topological features of intermediate configurations. Furthermore,<br>we construct a quantitative model based on the accumulation bar length A1 to predict the energy and<br>stability of protein configurations, which establishes a solid topology-function relationship of proteins with<br>the constant velocity pulling algorithm of SMD.