BAM + CHAIN = BAM? (liftover woes)

[Ajb123]

Hi SEQ Answers,

Long time reader, first time poster, all around noobie. I have been working with .bam files in a non-canonical reference (but have a .chain file linking the non-canonical reference back to b37). I want to create a new bam file that is based on b37 coordinates. Given that these files, what would be the fastest/most elegant way to convert/liftover to coordinates? Any applications you recommend specifically?

Best,

AJB

[Richard Finney]

You certainly can convert from bam to sam to bed-like then liftover then convert back from bedlike to sam to bam.

but ... Why not just dump the reads to fastq and re-align these reads to the "b37" genome? Probably the same level of work and you don't have to deal with the "umapped" (or "un-lifted-over") file.

[Ajb123]

Excellent! I just tried this protocol and everything seems fine (the reads are well positioned on their new reference), but they are missing their base calls and CIGAR string! The bed file doesn't seem to contain any of the short read bases or CIGAR string or quality scores. What should I do to preserve this information in the conversion? Thanks so much for your help and consideration.

I used BEDTools in the following manner:

bamToBed mybam.bam > mybed.bed
liftOver mybed.bed mychain.chain output.bed unmapped
bedToBam -i output.bed -g genome.g > output.bam