|
1. Positive dataset: 5227 domain pairs, extracted from 3DID database, were used as positive dataset in this study.
2. Negative dataset: 4089 domain pairs, extracted from PDB database , were used as negative dataset in this study. The strategy behind negative dataset creation is depicted in algorithm section of this site.
3. Independent dataset: 881 domain pairs, extracted from iPFAM database , were used as independent or blind set in this study.
Following are the domain pairs predicted as interacting by various existing methods-
1. SLA50_ddi.txt 1222 domain pairs taken from Jothi et al. 2006 J Mol Biol 362: 861-875.
2. SLA75_ddi.txt 392 domain pairs taken from Jothi et al. 2006 J Mol Biol 362: 861-875.
3. Lee at. al.,domain pairs 2391 domain pairs taken from Lee et al. 2006 BMC Bioinformatics 7: 269.
4. s1.txt 3000 domain pairs taken from Guimaraes et al. 2006 Genome Biol 7: R104.
5. s2.txt 3000 domain pairs taken from Guimaraes et al. 2006 Genome Biol 7: R104.
6. s3.txt 3610 domain pairs taken from Guimaraes et al. 2006 Genome Biol 7: R104.
7. s4.txt 3944 domain pairs taken from Guimaraes et al. 2006 Genome Biol 7: R104.
8. DPEA results 3005 domain pairs taken from Riley et al. 2005 Genome Biol 6: R89. Only 1812 domain pairs (pfam-A) are considered by us because rest of the pairs consisted atleast one pfam-B domain.
9. Guda et al., 65515 domain pairs taken from Guda et al. 2009 PLoS One 4: e5096.
| 
