Allele frequency table and pie chart discrepancy

[dmurphy4788]

Hi,
I am a little confused by how the % of modified reads is being calculated. Frequently I see that the pie chart indicates a lower percentage of modified reads than the alleles frequency table might suggest-(I've tried to attach an image for an example).
Is this because the different alleles shown in the frequency table contain insertions/deletions that are not near the expected cleavage position, and therefore are not recognized as modified reads?

[Colelyman]

Hi Dan,

Thanks for using CRISPResso and for your question. You are correct, in order for a read to be counted as modified (and thereby increase the percentage in the pie chart) a modification needs to occur within what we call the quantification window. By default, the quantification window is centered at the -3 position from the end of your sgRNA and is 1 nucleotide wide on either side. You can change where the quantification window is centered with the -wc or --quantification_window_center and how wide it is with the -w or --quantification_window_size parameters.

In this example, assuming that your quantification window size is 1, none of these most frequently occurring alleles shown will counted as modified because they all have an A to the left of the window center and a G to the right of the window center.

Hope that helps!

Cheers,
Cole

[dmurphy4788]

Hi Cole,
Great, thank you. That is just the information I was looking for.
One follow up question. I've noticed that the quantification window is often in the wrong place when the guide RNA targets the - strand. As in the example above, the PAM of the gRNA is the CCT sequence at the far left. Am I correct in thinking that the quantification window center then be centered on the G, as in 'CCTCTG|AA'. This would make the % of modified reads in the pie chart much more closely align with the results in the allele frequency table.
If so, would the best solution be to just flip the amplicon in the input? or manually change the window center?

Thanks,
Dan

[kclem]

Hi Dan,

If you see this discrepancy on the - strand, it's a bug and should be fixed. Do you have an example?

The gRNA should always be provided as the sequence immediately adjacent but not including the PAM sequence (5' of NGG for Cas9), and the cleavage offset is always relative to the end of the guide sequence. If the guide is found in the reverse complement of the amplicon sequence, the guide will be reversed (and the position of the predicted cleavage site is adjusted).

What was the command you used to produce the above output?

[dmurphy4788]

Hi,
Sorry for the delay.
I re-ran the code and double checked the orientation of the guides. I hadn't realized it but it looks like the guide sequences were reverse complemented at some point, so the input I gave was the plus strand, even though the guide targets the minus strand. This caused the error with the quantification window. Re-running the same code with the correct gRNA solved the issue.
Sorry about the confusion. I love this tool and I appreciate your help!

Dan

[kclem]

Good to hear! Happy analyzing!