Weird translated refs between hg19 and hg38

Question

[andrewl]

I'm possibly doing something wrong but having problems converting from hg19 to hg38, where my reference call is changed (which changes the user call). I tracked down one example that is reproducible:

#CHROM POS RSID REF ALT QUAL ALLELES

8 129076572 rs13255292 C T . 0/1

(this is hg19 - in this case the user is C/T)

I use crossmap to convert to hg38:

CrossMap.py vcf hg19ToHg38.over.chain.gz test.vcf hg38.fa test-converted.vcf

I am given this output

#CHROM POS RSID REF ALT QUAL ALLELES

8 128064326 rs13255292 A T . 0/1

Reading as it this gives me a A/T for the user.

What am I doing wrong here?

[BioinfGuru]

Cant' provide an answer but the posts below may be of use. Cross species mapping is something I'm beginning to look into. Crosmap is the basis for the Ensembl API. So maybe try the Ensembl API to see if it returns the same issue?

Converting Genome Coordinates From One Genome Version To Another (Ucsc Liftover, Ncbi Remap, Ensembl Api)

Translating Genomic Locations Between Species

[finswimmer]

Hello,

it looks like you have a off-by-one problem in your vcf file. The position for rs13255292 should be 8:129076573 in hg19 and 8:128064327 in hg38.

How do you generate the vcf file?

fin swimmer