Ambiguous reads #132
Replies: 6 comments 1 reply
-
Hi @RoebideBruijn, Ambiguous reads are those where your patient-specific mutation is probably deleted so CRISPResso can't tell whether to align it to HDR or WT. Can you paste an example of read alignments in 1. and 2? |
Beta Was this translation helpful? Give feedback.
-
Hi, Hereby an example for Q1): WT seq: HDR seq: The first two substitutions are silent mutations to prevent recutting. The last one (third one) is the one that is the substituion of interest (patient-mutation). It turns out we have a sample that has hardly any reads with the patient-mutation but a lot that have the 2 silent mutations. CRISPResso2 aligns these reads to the HDR sequence. And the '-w' parameter is quite small so it is even quantified as 'unmodified'. Therefore in the standard setting we would conclude that we have a lot HDR sequences. This would be the wrong conclusion since we are only interested in how many reads we have that have the patient mutation. I changed the -w parameter to quantify these reads at least as 'modified HDR'. But the biologist would like these to be quantified as 'Reference' sequence. But I was unable to get that done. Since from the 3 locations it has 2 in common with the HDR sequence and 1 with the Reference. So it will never choose the 'Reference' sequence as the better match. Do you have any idea how we can better deal with this issue? |
Beta Was this translation helpful? Give feedback.
-
I'm sorry -- can you give me an example of a read that aligns to the incorrect amplicon using CRISPResso? I'd need:
Does that make sense? |
Beta Was this translation helpful? Give feedback.
-
For question 1: WT seq: HDR seq: read: read classified as HDR (understandable from an algorithm perspective because it has 2 out of 3 substitutions in common). Nevertheless, biologically undesirable (missing the patient mutation, only has the 2 silent mutations). |
Beta Was this translation helpful? Give feedback.
-
For Q2 my mistake, the reads were longer than I thought, so the deletions was on the HDR/WT difference, hence correctly ambiguous. |
Beta Was this translation helpful? Give feedback.
-
Hi @RoebideBruijn, following up on Q1, It seems like your collaborator would like partial-HDR (with the first two silent mutations) to be classified as Reference? I would probably push back on this, because they're likely partial HDR at least. In this case, I would provide the WT amplicon and the reference with only the patient-mutation (not the two silent mutations) (perhaps you could call this amplicon PM for patient-mutation). This way, only reads containing the patient mutation will be quantified as PM. By examining the output, you'll also be able to measure the substitution rates at the other two sites in the reads that were assigned 'PM' vs 'WT' Does that make sense? |
Beta Was this translation helpful? Give feedback.
-
Dear all,
I'm using CRISPResso2 on a CRISPR setup with a ssODN sequence. To quantify the resulting reads I use as input the WT and the HDR sequence. The HDR sequence has the patient-mutation we are interested in (can be substitution, deletion, insertion) plus often some substitutions (silent-mutations) to prevent recutting. We are running into the following issues:
Anyone any idea?
Beta Was this translation helpful? Give feedback.
All reactions