Prosecutor: parameter-free inference of gene function for prokaryotes using DNA microarray data, genomic context and multiple gene annotation sources

Table 1 Confirmed results from Prosecutor

Organism	gene	functional category	rank	auc	reference
Campylobacter jejuni	Cj0391c	Pathway flagellar assembly	2	0.76	[41, 42]
	Cj1242	GO:0003774 motor activity	3	0.75	[42]
	Cj1316c	GO:0019861 agellum	7	0.72	[43]
Escherichia coli	yncE	GO:0015343 siderophore-iron transmembrane transporter activity	7	0.96	[44]
	ybiX	UP:Enterobactin biosynthesis	1	0.99	[44]
	cho	GO:0009432 SOS response	17	0.92	[45]
	ybeD	GO:0051082 unfolded protein binding	1	0.78	[46]
	ulaC	GO:0019852 L-ascorbic acid metabolic process	1	0.76	[47]
	yciW	GO:0006534 cysteine metabolic process	10	0.82	[48]
Bacillus subtilis	ypbG	transcriptional module SigM	18	0.81	[49]
	ykuO	transcriptional module Fur	8	0.75	[50]
	yviF	GO:0006935 chemotaxis	37	0.90	[51]
	ylxF	Pathway Flagellar assembly	7	0.89	[52]
	yfnE	transcriptional module GerE	52	0.81	[53]
Streptomyces coelicolor	SCBAC28G1.05	PW:Biosynthesis of type II polyketide back- bone	1	0.99	[54]
	SCBAC28G1.07	PW:Biosynthesis of type II polyketide products	8	0.76	[54]
Vibroi cholera	VC1688	GO:0006826 iron ion transport	2	0.87	[55]
	VCA0216	GO:0019290 siderophore biosynthetic process	8	0.98	[55]
	VC1267	GO:0019290 siderophore biosynthetic process	6	0.98	[55]

Functional predictions identified by Prosecutor for several organisms that are confirmed using literature information. Gene: the gene for which a validated functional prediction with a functional category (column three) was found. Rank: the position of a gene in the prioritized list based on p-values. These p-values describe the functional prediction significance for every individual gene with a specific functional category. AUC: the expression coherence value for a functional category with respect to its own members. Notice the examples of genes for which a particular annotation is assigned rank 1. This means that this gene is more close associated with this functional category then any of the original known members of the category, indicating a very high confidence in the prediction.

ISSN: 1471-2164