The readout of CRISPResso2

[flp461]

Describe the bug
After I run CRISPResso2,


We could see the reads number of these two mutations, TCG>TCA & TCG>TAG, are 9998 and 9395, which are already more than the reads of in-frame mutations, 15949. Where should I classify these substitutions? in-frame mutations?
The help would be appreciated.

[kclem]

Hi @flp461,

Thanks for using CRISPResso. Indeed, if the coding sequence includes these bases, these should count as in-frame mutations. Are you using the --ignore_substitutions flag?

Would you be able to provide the files and command you used to run? Or at least your reference and coding sequences?

[flp461]

Hi @kclem，

Also thanks for your software! It's very import for my current project.
These bases are included in coding sequence, I didn't use --ignore_substitutions flag.
I just attach the fold for it. Inside you could find fastq files and html report.
240.zip

/opt/conda/bin/CRISPResso --fastq_r1 240-R2-day2_S5_L001_R1_001.fastq --fastq_r2 240-R2-day2_S5_L001_R2_001.fastq -g AGGTGGACATCTGGTCGGCT -a CTGGACACCTTCTCCGGCTTCAAGGTGGACATCTGGTCGGCTGGGGTCACCCTGTAAGTGCCCCGCCCCCCCGGGCACTCACCACACGCACACTCCGAGGGGCCTCTGCGTCTTGGGCAGCTGCCGGCCTGTGGGCGCAGGGCGTGGCCACCGGCCCAGACCCTCTCTGGCCACAGCCGCTAGGGGGTGCTTACTTTATGGAAATGTAACTCATACGGCAGATGGT -c CTGGACACCTTCTCCGGCTTCAAGGTGGACATCTGGTCGGCTGGGGTCACCCT --default_min_aln_score 80 -n 240-R2-day2 --debug

Best,
Yiyuan

[kclem]

Thanks @B-1991-ing and @flp461,

Briefly, the algorithm currently only considers edits in the quantification window and their effect on the coding sequence. I think we originally did this to try to exclude modifications outside the quantification window which would probably be noise or at least not originating from the CRISPR edit.

In your example, the quantification window is only 1bp surrounding the cut site and only includes the A->G mutation (9998 reads) and not the C->A mutation (9395 reads). You can see the quantification window in figure 2b outlined by the dotted rectangle.

You can change the size of the quantification window with the -w or --quantification_window_size parameter https://github.com/pinellolab/CRISPResso2#quantification-window-parameters
In your case, set this as -w 2 and you should catch these modifications, or you can disable this window with -w 0.

Let me know if that makes sense.

If you'd like, you can try to convince me that I should include all edits in the exon (even those away from the predicted cut site) and we can look into making this change.

[flp461]

Hi @kclem，

Thanks a lot! You solved this problem.

For my case, it will be better for me to change the quantification window every time according to different KI design.

In addition, if I want to catch all the all edits in the exon, what should I do? Is it just to disable this window with -w 0?

Best,
Yiyuan

[kclem]

Glad that helped!

Yes, to catch all edits in the exon, seq -w 0, as well as --exclude_bp_from_left 0 and --exclude_bp_from_right 0. However, be aware that apparent mutations that may not arise from genome editing may also be added in your quantification.