A domain-based approach to predict protein-protein interactions

BMC Bioinformatics

Table 1 Details of the yeast MIPS datasets used in optimization studies *

* Starting yeast dataset was obtained from Munich Information Center for Protein Sequences (MIPS) site [35] and it contained 8250 positive and ~2 million negative protein-protein interactions [12]. Retained interactions column report the number of entries for the sets after the original dataset is filtered according to the domain pairs included as optimization parameters. Further details can be found in the Methods section. ^a These are the closed set versions of the 867, 2466, and 5095 parameter inclusive sets. As explained in the Methods section, during filtering to obtain the closed PPI sets, occurrence of some of the domain pairs are nullified and these parameters cannot be truly optimized during the GA runs. So these closed sets are a subset of their corresponding inclusive sets.

ISSN: 1471-2105