Computational discovery of sense-antisense transcription in the human and mouse genomes

Table 3 Properties of overlapping pairs of transcripts

(a)	Genomic arrangement				Number of pairs
	Tail to tail (3' to 3')				134
	Head to head (5' to 5')				12
	Transcript starts in intron of the other transcript				3
	Transcript contained entirely within the other transcript				39
	Difficult to classify				29
(b)	Candidate pairs	S/H	S/NH	NS/H	NS/NH
	S/H	92	24	18	65
	S/NH	-	0	3	2
	NS/H	-	-	1	8
	NS/NH	-	-	-	4

(a) For each candidate, we characterized the genomic arrangement of the pair of transcripts relative to one another, based on the observed mappings of sense and antisense ESTs to genomic sequence. We utilized categories developed by Lehner et al. [14] to facilitate comparison. Categorizations of individual candidate pairs are provided in the additional data files. As ESTs represent fragments of the full transcripts, we cannot be conclusive about these categorizations until full-length RNA species are cloned. (b) The characteristics of each candidate pair with respect to observed coding potential (based on homologs in protein databases) and observed splicing were tabulated. A summary of the distribution of joint characteristics of each pair is presented here. S/H, splicing and protein homologies observed; NS/H, protein homologies observed, but splicing not observed; S/NH, splicing observed, but protein homologies not observed; NS/NH, neither splicing nor protein homologies observed. Categorizations of individual candidate pairs are provided in the additional data files. As ESTs represent fragments of the full transcripts, we cannot be conclusive about these categorizations until full-length RNA species are cloned.

ISSN: 1474-760X