Automatic classification of protein structures using low-dimensional structure space mappings

BMC Bioinformatics

Table 7 Comparison of SCOP superfamilies and hierarchical clusters.

Cluster mapping	Max deg.	Mean deg.	Std deg.	Mean sim.	Std sim.	Mapped clusters	Ratio-1
SCOP- Matt(Z)	171	*1.928*	6.927	*0.394*	0.309	*1180 : 612*	*0.928*
SCOP-Matt(S12)	22	2.453	*2.324*	0.285	0.224	1180 : 481	1.453
SCOP-Dali(S12)	51	2.987	3.920	0.249	*0.219*	1180 : 395	1.987
SCOP-Matt(S24)	29	2.370	3.121	0.312	0.268	1180 : 498	1.369
SCOP-Dali(S30)	52	2.582	3.613	0.294	0.255	1180 : 457	1.582
Matt(Z)- SCOP	35	1.262	1.597	*0.492*	0.319	766 : 607	0.262
Matt(S12)-SCOP	22	1.180	1.074	0.400	0.251	564 : 478	0.180
Dali(S12)-SCOP	9	*1.082*	*0.521*	0.419	*0.247*	*422 : 390*	*0.082*
Matt(S24)-SCOP	19	1.178	0.942	0.464	0.280	590 : 501	0.178
Dali(S30)-SCOP	20	1.101	0.944	0.467	0.272	502 : 456	0.101

The best results for each mapping direction and column are highlighted in bold italics. Max, mean and std. degree refer to the number of Superfamilies or clusters corresponding to a given superfamily or group. Mean and std. similarity refer to the Jaccard coefficients obtained between these matching clusters. Finally, "mapped clusters" indicates the number of clusters or Superfamilies found, and the number of unique groups which to which are mapped. This ratio minus one is shown in the rightmost column; the closer this number is to zero, the more similar are aggregate sizes of the two tessellations. The values of min degree and max similarity are both one and the value of min similarity is zero.

ISSN: 1471-2105