Variable calling of m6A and associated features in databases: a guide for end-users.

N6-methyladenosine (m$^{6}$A) is a widely-studied methylation to messenger RNAs, which has been linked to diverse cellular processes and human diseases. Numerous databases that collate m$^{6}$A profiles of distinct cell types have been created to facilitate quick and easy mining of m$^{6}$A signatures associated with cell-specific phenotypes. However, these databases contain inherent complexities that have not been explicitly reported, which may lead to inaccurate identification and interpretation of m$^{6}$A-associated biology by end-users who are unaware of them. Here, we review various m$^{6}$A-related databases, and highlight several critical matters. In particular, differences in peak-calling pipelines across databases drive substantial variability in both peak number and coordinates with only moderate reproducibility, and the inclusion of peak calls from early m$^{6}$A sequencing protocols may lead to the reporting of false positives or negatives. The awareness of these matters will help end-users avoid the inclusion of potentially unreliable data in their studies and better utilize m$^{6}$A databases to derive biologically meaningful results.