Skip to main content
Figure 3 | BMC Bioinformatics

Figure 3

From: Application of nonnegative matrix factorization to improve profile-profile alignment features for fold recognition and remote homolog detection

Figure 3

ROC scores of various methods for remote homolog detection at the superfamily level. The performance of NMF features is overall better than that of the original PPA features. NMF features achieve over 0.99 ROC scores for 59% of all proteins in the testing set, whereas those of original PPA features, HHsearch, and PSI-BLAST are 30%, 43%, and 11%, respectively. The mean ROC scores of NMF features, original PPA, HHsearch, and PSI-BLAST are 0.95, 0.87, 0.93, and 0.75, respectively.

Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us