Neighbor effect and local conformation in protein structures.

In order to determine the preference or avoidance of the first and second positions of individual dipeptides for adopting different structural conformations, we randomly select defined structural groups of proteins from protein data bank and statistically analyzed the distribution of all 400 possible dipeptides in different secondary structural elements. Considering different combinations of α-helix (α), β strand (β) and coil (c) including αα, αβ, αc, ββ, βα, βc, cc, cα, cβ conformations, we found that some dipeptides are randomly distributed in these conformations, while others have non-random distribution for a given conformation. Finally, we provide new set of data containing preference and avoidance tendencies that originate from the neighbor effect for each amino acid according to the context of secondary structural element. The output of current work can provide novel data for different fields of structural bioinformatics as well as experiments involving site-directed mutagenesis.