Cas9 finds, then cuts both strands of DNA. Now you have to flat edges of a braid you must rejoin. This is called Non-homologous end-joining of DNA (NHEJ). The recognition, repair, fidelity, and correct repair during NHEJ is not well understood. Cas9 does no joining, no ligation, no pasting, it only does the snipping. The joining is done 'on it's own' by the cell at some rate, sometimes correctly, sometimes even uptaking an insert into the process. So insofar as Cas9 is doing any 'editing', the repair process is entirely an accidental after-effect of Cas9's targeted DNA breakage.

Here is CRISPR breaking DNA and creating the insult: https://www.ncbi.nlm.nih.gov/pubmed/27866150

Some recent work on understanding how those double-stranded breaks actually repair themselves: https://www.ncbi.nlm.nih.gov/pubmed/27924007

Some basic research in Yeast: https://www.ncbi.nlm.nih.gov/pubmed/27915381

Double-stranded break repair in breast cancer: https://www.ncbi.nlm.nih.gov/pubmed/28053956

I quickly looked through your links. The first article you link seems to lack any control group (which we need to assess the proportion of pre-existing mutant cells), the second is primary research but does not mention CRISPR-Cas9, and the other two are review articles that don't contain any quantitative data.

As mentioned, I only looked quickly, but I do not think any will be helpful on the point I am bringing up. Let me know if I missed it.

If you want to know why people are assuming NHEJ occurs at a double-stranded break, you might look at the dozen or so references in the first paragraph of https://www.ncbi.nlm.nih.gov/pmc/articles/PMC231204/pdf/1621...

Frequency of survival is shown in table 2... it ranges from 5/1,000 to 1/100,000. So cell death is a 200-100,000 times more common result than NHEJ using those methods. I have no idea how well that would correspond to the conditions we are interested here, but see no way this paper can be taken to support NHEJ is a more likely result than cell death.

It also says: "Yeasts, like other eucaryotes, exhibit cell cycle arrest in response to DNA damage (17)."

Right, so you are just saying no one really understands how NHEJ works at this point.

Have you considered that mechanism (NHEJ) is not even necessary to explain the rise in proportion of mutant cells after the CRISPR-Cas9 treatment? Because while it may happen to some extent, I do not think that is the simplest explanation for what I have seen reported.

I've been thinking this for awhile now, but never bothered to really write it up and keep track... you can check the discussion here (there are some other posts about CRISPR-Cas9 strewn about that thread as well): https://news.ycombinator.com/item?id=14014098

I have really only read papers other people bring up in conversations online or that get media coverage I come across on this topic. But every single paper I looked into has reported results that fit my explanation just as well as , or even better than, the NHEJ one, and this has been a couple years now...