Automatic structure classification of small proteins using random forest

BMC Bioinformatics

Table 4 Classification of NA-pairs in strategies 3 and 4 according to the probability estimates

Predicted Shared Level	Total	p < 0.5	p = 0.5	0.5 < p < 0.9	p ≥ 0.9
	Strategy 3
Class	2806	14	130	1235	1427
Fold	29	0	6	9	14
Super-family	2	2	0	0	0
Family	21	0	8	12	1
	Strategy 4
Class	4558	37	162	1917	2442
Fold	55	0	5	37	13
Super-family	3	3	0	0	0
Family	37	0	9	20	8

Following strategies 3 and 4, NA-pairs can be classified as sharing one of the top four SCOP structural levels. A large fraction of NA-pairs are classified to share the same Class in both the strategies. However, further distribution of such pairs according to the probability reflects the classification confidence. The high probability reflects more confidence in the predicted classification.

ISSN: 1471-2105