Question

Missing KRAS in locateVariants or TxDb.Hsapiens.UCSC.hg19.knownGene?

1

Entering edit mode

ckr123_tw ▴ 10

@ckr123_tw-6876

Last seen 10.2 years ago

United States

Hi,

While mapping mutations to hg19 known genes using locateVariants, several coordinates should be around KRAS (GeneID: 3845), but were not annotated. Am I missing something?

KRAS (uc001rgr.3)   chr12   25386768   25403863   transcript variant b
KRAS (uc001rgq.1)   chr12   25358180   25403854   transcript variant b
KRAS (uc001rgp.1)   chr12   25358180   25403854   transcript variant a

> table1
CHR STR END STRAND
1 chr12 25360414 25360414 +
2 chr12 25378561 25378561 +
3 chr12 25378647 25378647 +
4 chr12 25380275 25380275 +
5 chr12 25398284 25398284 +
6 chr12 25398285 25398285 +

Codes:
txdb <- TxDb.Hsapiens.UCSC.hg19.knownGene
MutC_GR = GRanges(seqnames = table1[,1], ranges = IRanges(start=table1[,2],end=table1[,3]),strand = table1[,4])
loc_all <- locateVariants(MutC_GR, txdb, AllVariants())
loc_all_table <- as.data.frame(loc_all)

> loc_all_table
seqnames start end width strand LOCATION LOCSTART LOCEND QUERYID TXID CDSID GENEID PRECEDEID FOLLOWID
1 chr12 25360414 25360414 1 + intergenic NA NA 1 NA NA <NA> 11228, 8082 10012632....
2 chr12 25378561 25378561 1 + intergenic NA NA 2 NA NA <NA> 11228, 8082 144363, 4033
3 chr12 25378647 25378647 1 + intergenic NA NA 3 NA NA <NA> 11228, 8082 144363, 4033
4 chr12 25380275 25380275 1 + intergenic NA NA 4 NA NA <NA> 11228, 8082 144363, 4033
5 chr12 25398284 25398284 1 + promoter NA NA 5 45946 NA <NA>
6 chr12 25398284 25398284 1 + intergenic NA NA 5 NA NA <NA> 11228, 8082 144363, 4033
7 chr12 25398285 25398285 1 + promoter NA NA 6 45946 NA <NA>
8 chr12 25398285 25398285 1 + intergenic NA NA 6 NA NA <NA> 11228, 8082 144363, 4033

software error annotation • 1.4k views

ADD COMMENT • link updated 10.2 years ago by Valerie Obenchain ★ 6.8k • written 10.2 years ago by ckr123_tw ▴ 10

score 2 · Answer 1 · 2014-10-16

2

Entering edit mode

Hervé Pagès 16k

@herve-pages-1542

Last seen 10 hours ago

Seattle, WA, United States

Hi ckr123_tw,

If you look at the man page for locateVariants(), you'll see that the method for signature GRanges,TxDb,VariantType has an ignore.strand argument, which by default is set to FALSE (questionable design choice given that variants are not strand-specific events):

ignore.strand: A ‘logical’ indicating if strand should be igored when
performing overlaps.

Try again with ignore.strand=TRUE.

H.

ADD COMMENT • link 10.2 years ago Hervé Pagès 16k

0

Entering edit mode

Thank you, Hervé Pagès. The ignore.strand=TRUE worked perfectly.

ADD REPLY • link 10.2 years ago ckr123_tw ▴ 10

1

Entering edit mode

Glad it worked. And as you can see, it worked for Val too because she didn't set the strand of her variants, which is the right thing to do. In your case you needed to use ignore.strand=TRUE because you set the strand of your variants to +, but the KRAS gene is on the minus strand. So maybe it's fine that locateVariants() is strand specific by default after all. It assumes that, unless the user has a good reason to do otherwise, the query will be unstranded (i.e. strand set to "*").

Cheers,

H.

ADD REPLY • link 10.2 years ago Hervé Pagès 16k

score 2 · Answer 2 · 2014-10-16

Hi,

I do get gene '3845'. What version of VariantAnnotation are you using? When posting a question please show the output from SessionInfo().

This was run with the current release version 1.12.0.

gr <- GRanges("chr12", IRanges(c(25360414, 25378561,25378647,
              25380275, 25398284, 25398285), c(25360414, 25378561,
              25378647, 25380275, 25398284, 25398285)))

library(TxDb.Hsapiens.UCSC.hg19.knownGene)
txdb <- TxDb.Hsapiens.UCSC.hg19.knownGene
loc <- locateVariants(gr, txdb, AllVariants())

> unique(loc$GENEID)
[1] "3845" NA

The current release and devel versions can be seen here:

http://www.bioconductor.org/checkResults/release/bioc-LATEST/

http://www.bioconductor.org/checkResults/devel/bioc-LATEST/

If your version is out of date, guidelines for updating are here:

http://www.bioconductor.org/install/

Valerie