Protein cutoff scanning: aplicação da varredura exaustiva de distâncias inter-resíduos na análise de contatos intracadeia em proteínas globulares

Abstract

In this study we carried out a comparative analysis between two classical methodologies used to prospect residue contacts in proteins: the traditional cutoff dependent (CD) approach and the cutoff free Delaunay tessellation (DT). Additionally, two alternative coarse-grained forms to represent protein residues were tested: using alpha carbon (CA) and using side chain geometric center (GC). A database was built, comprising two top classes according to CATH classification: all alpha and all beta. We found that the cutoff value at about 7.0 Å emerges asan important distance parameter in analysis of contacts in proteins. This value was not only independent of residue representation and of protein class but it was also the point where CD and DT methods diverged regarding their results. Up to 7.0 Å, CD and DT properties areunified, which implies that at this distance all identified contacts (edges) are fully truepositives (complete and not occluded). This unification may also imply that the edges distribution up to 7.0 Å is constituted mainly by contacts involving buried sites of the first coordination shell. We also have shown that DT techniques have a known anomaly, comprehending points near the degenerate condition, which in proteins may producedangerous and systematic errors affecting mainly the contact network in beta chains with CA residue representation. The almost-Delaunay (AD) approach has been proposed to solve this DT anomaly. We found that even AD may not be an advantageous solution. We empiricallydemonstrated that the DT+A results converge to CD, as the AD threshold perturbation parameter grows. This warns that DT and correlated techniques should be used with care in contacts analysis of proteins. As a consequence, in the strict range up to 7.0 Å, the CD approach revealed to be a simpler, more complete and reliable technique than DT (or DT+AD) to prospect protein contacts. Finally, we have shown that coarse-grained residue representation may introduce bias in the analysis of neighbors in cutoffs up to 6.8 Å, with CA in favor of all alpha proteins and GC in favor of all beta proteins. Beyond 6.8 Å, this bias isapparently eliminated. This provides an additional argument in benefice of the value 7.0 Å as an important lower bound cutoff to be used in contact analysis of proteins, for both CA and GC coarse-grained models.