Prediction of regulatory elements in mammalian genomes using chromatin signatures

BMC Bioinformatics

Table 7 Predicted active and inactive promoters in the mouse genome.

Active Promoter^a
Cell lines	Active Gene	Total Prediction	Refseq Supported PPV	Predicted promoters not present in the expression measurement	Expression Supported PPV
ES	7887	13853	81.4%	7191	88.6%
MEF	8092	11913	88.1%	5480	92.3%
NPC	7413	12700	84.1%	6259	89.0%
Inactive Promoter^b
Cell lines	Inactive Gene	Total Prediction	Refseq Supported PPV	Predicted promoters not present in the expression measurement	Expression Supported PPV
ES	4753	2862	77.0%	1806	79.2%
MEF	4248	4301	66.1%	3061	74.6%
NPC	5259	422	73.2%	267	94.8%

^a Active promoter supported by gene expression and Refseq: TP is the number of active genes prediction as active and FP is the number of inactive genes predicted as active. ^b Inactive promoter supported by gene expression: TP is the number of inactive genes predicted as inactive and FP is the number of active genes predicted as inactive. In both cases, we only considered predictions located within 2.5 Kb to annotated genes and the total number of predictions is thus usually larger than the sum of TPs and FPs. Refseq supported PPV shows how much percent of the total active/inactive promoter predictions are supported by Refseq.

ISSN: 1471-2105