A new pairwise kernel for biological network inference with support vector machines

BMC Bioinformatics

Table 2 Performance on reconstruction of the yeast co-complex networks.

	MLPK		TPPK		MLPK + TPPK		Direct
Data	Accuracy	AUC	Accuracy	AUC	Accuracy	AUC	AUC
Localization	76.2 ± 1.0	76.9 ± 2.0	79.5 ± 1.8	82.9 ± 1.7	80.6 ± 0.7	83.0 ± 1.2	73.9 ± 1.4
Chip-chip	82.2 ± 1.1	89.7 ± 0.8	63.8 ± 1.2	68.0 ± 1.1	84.4 ± 1.2	90.8 ± 1.2	58.4 ± 1.5
Pfam	92.1 ± 0.9	98.0 ± 0.5	86.1 ± 1.0	91.8 ± 0.9	93.8 ± 0.3	98.5 ± 0.1	67.3 ± 1.2
PSI-BLAST	89.0 ± 0.9	97.0 ± 0.1	88.3 ± 1.0	93.5 ± 0.9	93.1 ± 0.6	97.9 ± 0.2	67.8 ± 1.2
Sum	93.6 ± 0.3	98.7 ± 0.2	94.1 ± 0.6	98.0 ± 0.3	95.8 ± 0.3	99.1 ± 0.3	79.9 ± 0.8
Pairwise sum	93.3 ± 0.8	98.2 ± 0.4	90.5 ± 0.9	96.3 ± 0.7	95.2 ± 0.3	98.9 ± 0.2	-

The table lists, with the notation conventions explained in Figure 1, the results of the different methods on the reconstruction of the yeast co-complex networks.

ISSN: 1471-2105