Using Crispresso2 with gene family members

[faezehma]

Hi,
We are trying to use Crispresso2 for characterization of genome editing efficiencies of our target gene, G1. The primers we designed on the target site of G1 inevitably also pick up another gene family member, G2, elsewhere in the genome. Luckily, G1 and G2 do have a few SNPs within the amplicon. Easily up to 30% of the total reads go to G2, but this may vary from sample to sample. G2 however is of no interest to us, but with its high percentage of reads it will skew the editing percentages of our target gene G1. So, what we need is a way to somehow get rid of the reads mapping to G2 or isolate only the reads mapping to G1. However, Crispresso2 does not seem to have a differential mapping step.
I can say this is an issue when dealing with a genome with subgenomes, where we actually want separate results for each subgenomes.
do you have any suggestion how to deal with this issue in Crispresso2?

Another question, if Crispresso uses q scores only for the initial read filtering/processing or also for mapping and creating alleles tables?
Thanks!

[kclem]

Hi @faezehma,

I'd suggest you provide both amplicons to the analysis. Reads will be aligned to each amplicon separately (G1 vs G2) and quantification of editing will be performed and reported separately as well. Provide the amplicon sequences as a comma-separated list like this -a AAATCC,ATATCC or in the Amplicon box on the website.

You can also provide amplicon names as a comma-separated list like this -an G1,G2 or in the Amplicon Name/s box on the website.

Here's an example of running with multiple amplicons.

Note that if indels have deleted any of the SNPs that distinguish G1/G2 the reads will be classified as 'ambiguous' because they can't be assigned uniquely to either amplicon.

For your second question, q scores are mostly used for filtering and not for mapping. However, you can replace low-quality bases with 'N' which will align equally well to any base in the alignment. You can perform this using the --min_bp_quality_or_n parameter.

Let us know if that doesn't work!