Table I. Summary of previously existing subcellular annotation and comparison with STEPdb. Uniprot and EchoLOCATION databases contributed 2050 and 3979 protein annotations correspondingly and together with the theoretical IM proteome (33) gave a combined 4111 initially annotated proteins. Comparison of the three resources revealed some matches in proposed subcellular locations (“Matching annotation”) but also some differences (“Conflicting annotation”). IM proteins with matching annotations in only in two of three resources are referred as “Unique IM proteome.” All proteins with existing annotations in only one source are referred as “Unique (total).” STEPdb multicombinatorial analysis contributed 36 newly classified proteins that were of unknown location (“STEPdb de novo annotated”), 674 proteins that have been reassigned to locations other than previously proposed (“STEPdb revised”) and 601 proteins with contradicting subcellular annotations that have been unresolved (“STEPdb unresolved”).

	Uniprot	EchoLOCATION	Bernsel & Daley (2009)33	Total
Reference proteome (E. coli K-12)	4303	4345a	1133	4303
Matching annotations	1613	1646	850	1652
Unique (IM proteome)	11	29	4	44
Unique (total)	12	1998	4	2014
Contradicting annotations	425	599	254	601
Existing annotations	2050	4243	1108	4267
% of reference proteome	48%	98%	26%	99%
Missing annotations	2253	60	–	36
Missing and unresolved annotations	2678	659	254	637
% of reference proteome	62%	15%	6%	18%
STEPdb total contribution over previous annotations	3352	1333	560	1311
de novo annotated	2253	60	84	36
Revised	674	674	222	674
Resolved	425	599	254	601
Experimental validations				1205
References added				118
% of reference proteome	76.89%	35.37%	7.87%	32.81%

^a This is the estimation of the total proteome of EchoLOCATION.