Skip to main content
. 2019 Feb 14;47(7):e41. doi: 10.1093/nar/gkz074

Figure 1.

Figure 1.

Generation of positive and negative data. The transcriptome m6A sites under RRACH consensus motifs that have been reproduced in at least two of the five training datasets were used as positive m6A sites. The negative training data (non-m6A sites) was randomly selected from the non-positive RRACH adenosines on the full transcripts containing the positive sites.