Position of insertion in Homopolymer

[tristan dubos]

Hi all ,

I m using bwa for alignment and Varsan with mpileup to find my variants positions. In this list I found an insertion of A in homopolymer of A and the position of this insertion is the beginning of the homopolymer. I compared this result with an other work (witch analyze data with other bioinformatics tools) with this data set and all of my snp agreed only the insertion of A in homopolymer of A disagreed. In this other work the position of this insertion is the end the homopolymer.
Do you know is there rules for this kind of variant ? Is this difference could come from aligners ?

king regards

Tristan

[m_two]

see https://www.biostars.org/p/66843/

[Brian Bushnell]

Different aligners treat homopolymer indels differently. BBMap always shifts them in the same direction (rightmost, I think). PacBio's BLASR places them seemingly at random within the homopolymer, though it is probably related to the specific quality scores of the bases.

[tristan dubos]

Thank you to you two ! I understand now the number of variants we can miss if we compare it into database (like COSMIC) and don't use the same way too call variants .....