Optimal transport technique to understand peptide conformations

Using Protein Data Bank (PDB), the distribution of backbone torsional angle for dipeptides and tripeptides has been obtained and analyzed in several past works. However, the more extended peptides distributions are difficult to obtain due to the data sparsity. In this paper, we introduce a technique based on Optimal Transport theory to understand peptide conformations. This technique renders a computational method to determine the multi-point distribution of the backbone torsional angles of the peptide. Applying this technique, we study tetrapeptide conformations, presenting a detailed analysis of tetrapeptides distributions composed of Ala and Gly amino acids, and compare it with PDB data. Our study shows that tetrapeptides such as AAAA, AAAG, AGAA, and GAAA prefer right-handed alpha-helix secondary structure formations and beta-turns in GGGG, GAGG, AAGG, and AAGA.