"Duplicate" kgID when doing annotation.

[sxwcasd]

I am trying to running the annotation (svaba-annotate.R) on GENCODE db.
However, the UCSC db records pulled by these 2 lines.

svaba/R/svaba-annotate.R

Lines 59 to 60 in 0f60e36

	genes <- suppressWarnings(data.table::as.data.table(query(paste0("SELECT name, chrom, txStart, txEnd, strand, exonStarts, exonEnds, exonCount FROM ", assembly, ".knownGene"))))
	codes <- suppressWarnings(data.table::as.data.table(query(paste0("SELECT kgID, mRNA, geneSymbol, spID, refSeq FROM ", assembly, ".kgXref"))))

Are having duplicates. example:

                   kgID         mRNA      geneSymbol   spID       refSeq chrom   txStart     txEnd strand
  1: ENST00000244174.11    NM_002186            IL9R Q01113    NM_002186  chrX 155997695 156010817      +
  2: ENST00000244174.11    NM_002186            IL9R Q01113    NM_002186  chrY  57184215  57197337      +

This is making sense to me, that the sex chromosomes have different position and share the some mRNA. But this will hit error at the following line:
Error: !any(duplicated(genes$kgID)) is not TRUE

Maybe we have have a better validation check at here?

[walaj]

Hmm, that is an unusual situation with a gene listed on multiple chromosomes -- interesting catch. Happy to entertain a pull request if you have a better validation scheme.

…

On Thu, Jun 23, 2022 at 4:37 PM Linghao Song ***@***.***> wrote: I am trying to running the annotation (svaba-annotate.R) on GENCODE db. However, the UCSC db records pulled by these 2 lines. https://github.com/walaj/svaba/blob/0f60e366c300bbefbba762bcc6d2b661bd2ae74a/R/svaba-annotate.R#L59-L60 Are having duplicates. example: kgID mRNA geneSymbol spID refSeq chrom txStart txEnd strand 1: ENST00000244174.11 NM_002186 IL9R Q01113 NM_002186 chrX 155997695 156010817 + 2: ENST00000244174.11 NM_002186 IL9R Q01113 NM_002186 chrY 57184215 57197337 + This is making sense to me, that the sex chromosomes have different position and share the some mRNA. But this will hit error at the following line: Error: !any(duplicated(genes$kgID)) is not TRUE <https://github.com/walaj/svaba/blob/0f60e366c300bbefbba762bcc6d2b661bd2ae74a/R/svaba-annotate.R#L67> Maybe we have have a better validation check at here? — Reply to this email directly, view it on GitHub <#115>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABUZ7CG5OHOZAOFME3GMGODVQTDI7ANCNFSM5ZVTSDKQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[sxwcasd]

Because we need both gencode and exonframe information, we loaded wgEncodeGencodeCompV36 table instead of knownGene or kgXref. And validate by start and end sites. I m not sure if that make sense to your original design (because we don't need to consider refseq id mapping). But it seems accomplished our goals and get around this issue.