Advice on dealing with edited genome

[Matt K]

Hi,

First, big thanks to Alex and the group for helping me over the years!

I have a question on how to deal with an edited genome. I have a mouse gene that has an exon replaced with the human version. The human exon is slightly shorter than the mouse exon, so if I replace the endogenous genomic sequence with the edited sequence, the coordinates in the gtf for all subsequent genes will be wrong. 

My approach was:

1) Add a new chromosome to the fasta that represents the edited gene loci sequence and add a corresponding entry to the gtf for this new chromosome/gene

2) Remove the endogenous gene from the gtf to avoid multiple mapping issues

My questions are:

1) The endogenous sequence is still present, but not associated with a gtf entry. Will this still be flagged as a multiple mapper and not added to the count table, or would this work correctly?

2) More generally, is there an established way to created references from edited genomes? I imagine this is fairly prevalent now, but couldn't find any advice online.

Thanks for your time!

Best,

Matt

[Alexander Dobin]

Hi Matt,

1) The endogenous sequence is still present, but not associated with a gtf entry. Will this still be flagged as a multiple mapper and not added to the count table, or would this work correctly?

For --quantMode GeneCounts option it will be considered a multimapper and not counted.

2) More generally, is there an established way to created references from edited genomes? I imagine this is fairly prevalent now, but couldn't find any advice online.

 

I am not aware of an established way to do it. I would recommend masking the entire endogenous gene locus (exons and introns) with Ns, adding the edited gene locus. You will need to transfer (and edit) all the annotations (exon lines).