Conserved residue clusters at protein-protein interfaces and their use in binding site identification

BMC Bioinformatics

Table 2 Statistics showing the significance of clustering of conserved interface residues compared to the clustering in the subsets of the same-size containing randomly selected interface residues from the same structure

Datasets	Selection of conserved interface residues^a
	s < < s> _int				s < (< s> _int - σ)
	Number	M _s,cons	< M _s,random >	P -value ^b	Number	M _s,cons	< M _s,random >	P -value ^b
Homodimers	121	0.081 (0.02)	0.071 (0.02)	1.6E-04	103	0.087 (0.02)	0.070 (0.02)	2.0E-08
Complexes	389	0.103 (0.03)	0.089 (0.02)	6.5E-14	309	0.113 (0.04)	0.090 (0.02)	2.2E-16

a Conserved interface residues were selected using different criteria (see Methods and Table 1 footnote).
b P-values refer to the significance levels for the Mann-Whitney U-test corresponding to the hypothesis that M_s,cons is greater than < M_s,random > (i.e., the degree of clustering of conserved residues within an interface is greater than that of random, same-size subsets collected from the same interface). A P-value < 0.01 indicates that the difference is significant at the 1% level. Calculated using R [64].

ISSN: 1471-2105