TESTLoc: protein subcellular localization prediction from EST data

BMC Bioinformatics

Table 1 Performance of available tools and TESTLoc on plant EST-peptides¹

Predictors	chl²		cyt		end		ext		mit		nuc		per		pla		vac
	SN	PPV	SN	PPV	SN	PPV	SN	PPV	SN	PPV	SN	PPV	SN	PPV	SN	PPV	SN	PPV
TargetP	19	62					29	25	18	44
Protein Prowler	14	66					34	32	26	60			25	100
BaCelLo	61	69	50	28			60	32	9	90	88	39
Wolf-PSORT	27	41	60	27	0	0	18	36	8	58	65	71	16	50	0	0	0	0
YLoc	25	77	84	13	22	13	52	98	35	84	82	80	50	26	0	0	15	31
KnowPred	NA	NA	71	23	0	0	23	36	61	49	86	40	58	54	0	0	NA	NA
MultiLoc2	14	81	84	12	9	8	44	48	18	37	48	67	67	10	0	0	45	54
BLAST³	76	97	77	62	64	100	81	95	57	87	77	96	25	60	0	0	0	0
TESTLoc	99	99	50	88	20	20	71	98	86	71	75	96	50	21	30	5	63	100

¹ Numbers in %. Bold numbers are the result of the here described TESTLoc, which is tailored for ESTs. The results of full-length proteins are compiled in Additional file 8
² Abbreviations: chl, chloroplast; cyt, cytosol; end, endoplasmatic reticulum; ext, extracellular space; mit, mitochondrion; nuc, nucleus; per, peroxisome; pla, plasma membrane; vac, vacuole; SN, sensitivity; PPV, positive predictive value
³ Note that all test data have homologs in databases, which in practice is rarely the case; see text

ISSN: 1471-2105