Assessing the precision of high-throughput computational and laboratory approaches for the genome-wide identification of protein subcellular localization in bacteria

Table 1 PSORTb v.2.0 predicted localization sites for 405 proteins reported in ten subproteome studies.

Laboratory Data			PSORTb v.2.0 Predicted Localization^a)
Organism	Fraction^a)	Total	C	C/CM	C/P	CM	CM/P	P	P/OM	OM	OM/EC	EC	UN	Agreement^b)	Coverage^c)
E. coli [64]	C	23	19	-	-	-	-	1	-	-	-	-	3	95.0	87.0
Synechocystis [15]	CM	63	13	2	-	5	1	6	-	5	1	-	30	24.2	52.4
Synechocystis [46]	P	57	2	-	1	-	-	8	-	3	-	-	43	64.3	24.6
K. pneumoniae [16]	OM	3	-	-	-	-	-	-	-	3	-	-	-	100.0	100.0
S. typhimurium [16]	OM	11	2	-	-	-	-	-	-	6	1	-	2	77.8	81.8
E. coli [17]	OM	39	3	-	-	-	1	3	-	22	1	1	8	74.2	79.5
P. gingivalis [18]	OM	6	-	-	-	-	-	-	-	2	-	1	3	66.7	50.0
P. aeruginosa [13]	OM	33	4	-	-	1	-	-	1	22	2	1	2	80.6	93.9
P. aeruginosa [13]	EC	150	33	-	-	5	1	33	-	9		6	63	6.9	58.0
H. pylori [19]	EC	20	3	-	-	-	-	2	-	4	1	1	9	18.2	55.0

a) C = cytoplasmic, CM = cytoplasmic membrane, P = periplasmic, OM = outer membrane, EC = extracellular, and UN = unknown.
b) Percentage of agreement is defined by , where: A represents the number of proteins of the fraction X predicted by PSORTb to be resident at X and X/Y localization sites.
B represents the total number of proteins of the fraction X predicted as not unknown by PSORTb.
c) Percentage of coverage is defined by , where: B represents the total number of proteins of the fraction X predicted as not unknown by PSORTb.
T represents the total number of proteins identified in the fraction X.

ISSN: 1471-2164