Integrative approaches to the prediction of protein functions based on the feature selection

BMC Bioinformatics

Table 6 Performance of GO term prediction of yeast genes

Specificity	# of GO terms	KLR			K L1LR				KLR with Relief
		All		ES	λ / λ _max = 0.01		λ / λ _max = 0.1
		NOS*	S*		NOS*	S*	NOS*	S*
3-10	567	0.51	0.51	0.61	0.68	0.70	0.67	0.68	0.52
11-30	348	0.70	0.70	0.79	0.80	0.82	0.78	0.79	0.66
31-100	210	0.75	0.75	0.80	0.79	0.82	0.77	0.79	0.64
101-300	121	0.77	0.77	0.79	0.77	0.80	0.73	0.78	0.65

GO terms are categorized into four groups based on the number of genes covering the GO term. Prediction quality is estimated using AUC values for KLR with all data sources, KLR with a data source selected by exhaustive search (ES), KL1LR method (with two regularization parameters), and KLR with features selected by the Relief method. NOS stands for non-standardization of the features, S for standardization.

ISSN: 1471-2105