AceView: a comprehensive cDNA-supported gene and transcripts annotation

Table 1 Statistics of the 25 selected tracks, arranged in the order of the UCSC genome browser

UCSC track	Model with introns	Model with introns and CDS	Single exon model (some clipped)	Unique introns in mRNA	All introns in mRNA	Input or method
HAVANA Gencode (Sanger, UK) known + putative	1,691	649	70	3,618	9,693	MEP,CA,H
EGASP model submissions
AceView (NCBI, US)	1,630	1,460	24	3,530	9,597	ME,(H)
UP Dogfish (Sanger, UK)	204	204	15	1,679	1,679	CA
Exogean (ENS, France)	554	538	2	2,855	6,178	MEP,CA
UP ExonHunter (U Waterloo, Canada)	807	807	220	3,237	3,237	MEP,CA
Fgenesh (U London, UK)	462	458	97	2,610	3,241	P,CA
UP GeneId (IMIM, Spain)	267	267	51	1,905	1,905	A
UP GeneMark (Georgia IT, US)	551	551	81	2,185	2,185	A
UP Jigsaw (TIGR, US)	259	259	67	2,168	2,168	MEP,CA
PairagonAny (Wash U, US)	471	437	38	2,300	3,470	MEP?,CA
UP SGP2 (IMIM, Spain)	552	552	159	2,645	2,645	P,CA
P Twinscan-MARS (Wash U,US)	547	547	108	2,501	4,943	CA
UP Augustus Any (U Göttingen, Germany)	312	316	87	2,291	2,291	MEP,CA
UP GeneZilla (TIGR, US)	477	477	179	2,758	2,758	A
UP Saga (UC Berkeley, US)	331	331	47	1,737	1,737	CA
UCSC gene tracks
*Known Gene (UCSC)	501	477	53	2,264	4,427	MP
*P CCDS	201	201	14	1,296	1,508	MP,H
*RefSeq (NCBI, US)	342	325	41	2,082	2,922	M(E)P,H
*MGC	323	310	19	1,400	2,101	M
*Ensembl (EBI, UK)	427	418	58	2,429	3,548	MEP,CA
*AceView (Aug 2005 NCBI)	1,792	1,627	902	3,812	9,792	ME, (H)
*ECgene (Korea)	3,851	3,551	2,569	3,942	30,660	ME,C
*U NscanEst (Wash U, US)	282	252	27	2,292	2,292	ME,CA
*UP GenScan (MIT, US)	395	395	59	3,042	3,042	A

The number of models, with or without introns (after clipping at region boundaries), the number of spliced coding models, and the number of unique and multiply used introns are given over the 31 ENCODE test regions. Coded information has been added in front of the track name: asterisks distinguish standard gene tracks, available genome-wide, from an ENCODE only track; a U track predicts a unique model per gene; P predicts protein coding regions only. According to their documentation, the programs use different input or methods: M, E, P stand for human mRNA, EST, protein sequences or alignments, respectively; C stands for for conservation, or use of cDNA or protein evidence from other species; A stands for ab initio prediction; H stands for Hand curation; and parenthesized letters stand for minimal use of the particular type. Notice the low proportion of Gencode mRNA models with an annotated CDS (in bold).

ISSN: 1474-760X