Combining in silico prediction and ribosome profiling in a genome-wide search for novel putatively coding sORFs

BMC Genomics

Table 2 Coding potential of sORFs in different genomic locations

Genomic location	# of sORFs^a	Coding sORFs^b	Pcod > 0.9^c	Pcod > 0.99^c	Ribo sORFs^d	Coding Ribo sORFs^e
ncRNA	20,810	9,922	6,443	1,100	528	401
Exonic	63,180	34,063	21,546	10,872
Other	155,633	80,891	37,730	9,894
Intronic	417,277	34,845	14,582	2,361
Intergenic	1,757,458	223,235	107,567	27,371	226	89

Number of sORFs divided per genomic region and for which certain in silico and/or expression evidence can be found. Included are total number of sOFs with high coding potential (according to sORFfinder), number of sORFs having scores above certain thresholds (according to SVM analysis), number of sORFs which show ribosome profiling expression and number of sORFs for which in silico coding as well as expression evidence is available.
^a Total number of sORFs with high coding potential according to sORFfinder.
^b Total number of sORFs classified as coding by SVM^light.
^c Pcod is the coding probability score as predicted by SVM^light.
^d sORFs with mapped ribosome profiles, attaining sequence read coverage > 75% of the total ORF (based on cycloheximide treatment), and ribosome profile hits at the ORF start site (based on harringtonine treated samples).
^e Ribo sORFs (see under ^d for description) classified as coding by SVM^light.

ISSN: 1471-2164