Parity in the number of atoms in residue composition in proteins and contact preferences

Abstract

Two ways of looking into the amino acid composition in a database - one based on counting residues and another on the number of atoms each residue contributes to the total - give quite different results. In the former, the composition is inversely proportional to the residue size - the maximum deviation shown by Cys and Leu can be explained by their chemical property or the role in oligomeric structure, while the latter is size-invariant. Similarly, calculations of contact preferences between residues can be atom-based or residue-based, and the former method provides values that are size-independent, and thus should be preferred in modelling and docking studies.