variant calling after reference genome is updated

Question

[prasundutta87]

Hi,

If I have already performed WGS variant calling using a reference genome and if an updated assembly is released, do we again start the alignment and variant calling process using the new genome? Can someone let me know what is the best way to handle such case? The updated reference may have new chromosome names or an updated assembly (like unplaced scafflods have been successfully assigned a chromosome in the new assembly). Kindly let me know if more information is required. Specifically, NCBI gave the chromosome names their own accession number and I have been using a copy of the genome that has the chromosome names as 1,2, 3 and so on.

Furthermore, whenever an updated assembly comes, what kind of checks should be performed to see if any minor or major change has happened (except genome genome alignment dotplot)?

[GenoMax]

if an updated assembly is released do we again start the alignment and variant calling process using the new genome?

If the sequence coordinates changed in that process then the answer is yes. No way around it. If only the names have been changed then you could update them accordingly everywhere.

[prasundutta87]

Thanks a lot genomax..what should be the best way to check if the coordinates are same? As of now, I just made a faidx index (.fai) and checked the lengths of the chromosomes from the old and the new genome assembly. They are the same. Is that the correct way to do such checking?

Answer 1

[finswimmer]

Hello,

repeating Variant Calling wouldn't be enough. You must also repeat the alignment. Using different references in this two steps can end up in chaos.

fin swimmer

[prasundutta87]

Thanks Fin. Sorry, I meant that as well. I have edited the question. But if only the chromosome name has been updated? I believe that only mapping the old names to the new name would be enough?

[finswimmer]

Why do you want to update the names? What benefit do you think you have with doing so?

[prasundutta87]

The thing is that I am using the copy of the genome before it was submitted to ncbi for annotation. Ncbi annotation pipeline changed the chromosome numbers and gave its own id. If I want to publish my work, I may be challenged for that. Furthermore, the gff file also has the new chromosome names. That may be an issue while variant annotation.