(a) Graphical representation of the miRNA seed prediction pipeline. Based on the defined 3′-UTR from “Genbank Refseq” gene annotation, in genome sequences, we employed a sliding window method that used continuous 7 nt window with a step size of 1 nt starting from the 5′ terminal of defined 3′-UTR and calculated the WCS by the average conservation score in each window. We discarded sliding windows overlapped with repeat sequence. Then, we ranked all of these WCSs within each species and calculated the counts of conservation scores greater or less than the score of CS50, respectively. At last, for each type of 7-mers, we compared these two kinds of counts in all species to check whether the difference between them is significant using the binomial test followed by Bonferroni post hoc correction. We kept the 7-mer nucleotides as candidate miRNA seed if the conservation signal above background with value < 0.05. (b) Venn plot of known seeds (left circle) from a previous study and detected seeds (right circle). (c) Venn plot of known seeds from eukaryote (left circle) and detected seeds (right circle).
