In silico miRNA prediction in metazoan genomes: balancing between sensitivity and specificity

BMC Genomics

Table 5 Mining the C. elegans genome for putative miRNA hairpins

	Rejected miRNA hairpins ^a	Clustered ^b (miRNAs/loci)		Similar ^b (miRNAs/loci)
5 kb flanking sequence	NA	132	15,514	-	-
Similarity to mature miRNA ^c	NA	-	-	132	937
No overlap with exons ^d	2	130	11,197	130	706
L = 1.0	87	45	80	-	-
L >= 1e-5	15	-	-	116	197
Exclude TRF overlap ^e	0	45	74	-	-
Filter on 7 descriptors	6	-	-	116	162
Exclude known miRNA loci	132	0	20	0	64
similarity positioned correctly ^f	18	-	-	0	41

^a Rejected C. elegans miRNA hairpins by this step alone
^b Remaining cumulative number of C. elegans miRNA hairpins; remaining cumulative number of genomic hairpin loci
^c Similarity to a metazoan mature miRNA of at least 19 nt length with at most three mismatches
^d Cel-mir-354 is fully located in the final exon of Y105E8A.16; cel-mir-356 is largely located in the 3'utr of ZK652.2, but overlaps 6 nt with the final exon
^e At most 40 nt overlap with a repeat identified by Tandem Repeats Finder [47]
^f Similarity with mature miRNA may have at most 3 nt overlap with hairpin loop coordinates and at least 8 nt of the stem must separate the similarity from the end of the hairpin. Criteria are set according to the biological model of miRNA maturation from hairpins [3].

ISSN: 1471-2164