I'm using DMRcate to identify and annotate differentially methylated regions and I think it's giving incorrect information about overlapping genes, at least for some non coding RNA. I even see this in the DMRcate pdf manual https://bioconductor.org/packages/release/bioc/html/DMRcate.html.

On pages 5/6 the first range is chr2:87014979-87021117, using genome hg19. It says it overlaps SNORA73. In the UCSC Genome Browser SNORA73 is at chr18:53,746,625-53,746,825.

The second range in the manual is chr17:47286445-47289036, it says it overlaps SNORA69 and the gene ABI3. ABI3 is correct but the UCSC Genome Browser says SNORA69 is at chrX:118,921,316-118,921,447.

The fourth range correctly shows the gene SKI at chr1:2160666-2166155 while the Browser lists the non-coding RNA snoU13 at 4 possible locations, none of which overlap this region.

I've check a few of the ranges from my own data, the genes are correct the snoRNAs aren't.

I did a fresh install of R, Bioconductor, and all my packages less than 2 weeks ago so everything should be up to date. I see the same results in data analyzed several months ago.

Could this be a bug in DMRcate? What database is DMRcate using for its genes, particularly non-coding RNAs?

Update: DMRcate uses R package DMRcatedata to get gene info, the DMRcatedata manual says it uses this database for hg19: ftp://ftp.ensembl.org/pub/release-75/gtf/homo_sapiens/Homo_sapiens.GRCh37.75.gtf.gz. I downloaded it and looked for all lines with SNORA73. None of them show SNORA73 at chr2:87014979-87021117 like DMRcate says it is, there's one on chr2 but it's at chr2:3628160-3628322 which matches a Gencode annotation in the UCSC Genome Browser.