Detection and characterization of 3D-signature phosphorylation site motifs and their contribution towards improved phosphorylation site prediction in proteins

Table 2 Prediction performance as measured by the AUC

Kinase family	Kinase group	N	Sequence-only	Spatial-information enriched	NetPhos 3.1b	DisPhos 1.3	KinasePhos 2.0*
Ser kinases	/	363	0.74 ± 0.02	0.79 ± 0.02	0.69 ± 0.02	0.73 ± 0.05	0.63 ± 0.05
PKA	AGC I	34	0.91 ± 0.04	0.94 ± 0.04	0.91 ± 0.03
PKC	AGC II	31	0.83 ± 0.05	0.87 ± 0.04	0.78 ± 0.05
MAPK	CMGC II	12	0.89 ± 0.07	0.91 ± 0.06	0.78 ± 0.09
CKII	CMGC IV	19	0.73 ± 0.07	0.78 ± 0.07	0.76 ± 0.07
Thr kinases	/	134	0.72 ± 0.03	0.74 ± 0.03	0.66 ± 0.03	0.72 ± 0.06	0.66 ± 0.05
Tyr kinases	/	253	0.69 ± 0.02	0.71 ± 0.02	0.65 ± 0.02	0.56 ± 0.06	0.54 ± 0.05
SRC	PTK I	24	0.72 ± 0.07	0.79 ± 0.06	0.62 ± 0.07
unspecific predictor		750	0.71 ± 0.01	0.75 ± 0.01	0.67 ± 0.01	0.68 ± 0.03	0.63 ± 0.03

Results from the cross-validation of the various prediction approaches. The sequence-only and Spatial-information enriched methods were developed as part of this study and compared to NetPhos 3.1b that includes the kinase-specific predictor NetPhos/K, DisPhos1.3 and KinasePhos2.0. As KinasePhos reports only decision values of positively predicted sites, the evaluation of kinase specific prediction was not possible due to missing score values for sites not predicted to be phosphorylated. However, the kinase-specific predictions were feasible as KinasePhos essentially reports all submitted sites as being phosphorylated by at least one kinase. For the evaluation of the predictor, the highest reported decision value was used for each site. Best performing methods are printed in bold-face.

ISSN: 1471-2105