Mathematical models of gene editing using CRISPR

Question

One confusing thing I have found when reading articles about possible CRISPR based gene therapy treatments in humans is that there is vey little discussion about what percentage of your cells will actually have their DNA edited, the rate at which the editing takes place, and how to quantitatively estimate how many edited cells would be needed to "cure" a specific disease. (It's not so confusing how edits could be made to a small embryo, what's confusing is what is actually possible for grown patients)

Can someone provide a few basic facts that calibrate what is currently possible and what is needed.

For instance, is it currently possible to edit more than 50% of the cells of an adult nematode (or more complicated organism) using CRISPR? How long does it take (minutes, hours, days, weeks).

For the current list of human diseases that are possible targets for gene therapy, how many cells need to be edited for one of those diseases to be "cured" (hundreds, thousands, billions, etc)

But in general is there a name for the mathematical models that try to predict how many cells will end up being edited, how those edited cells will multiply and replace diseased tissue, and how many edits are needed to generate a concentration of missing protein for a treatment to be successful?

cheers!

Answer 1

There are several probabilities that you need to take into account.

• The efficiency of DNA transformation of CRISPR encoding DNA and target DNA into the host cell (varies between cell types and transformation agent).
• Efficiency of cutting by CRISPR at target site. (varies by DNA compaction - which varies by site and cell type, and guide RNA)
• Number and cutting efficiency at off targeting sites.
• Efficiency of degradation of free DNA by cellular antiviral systems (varies between cell types)
• Efficiency of DNA repair machinery to fix dsDNA breaks. (thus removing CRISPR cut sites. Varies by cell type)
• Efficiency of recombination machinery (homologous recombination and microhomologous recombination) (which is used to insert new DNA sequence).
• Probability of NHEJ damaging target site. (which can delete the CRISPR cut site)

Too many things to count.

However to give you some idea. The efficiency of CRISPR making a correctly targeted knock in for DT40 (chicken lymphoma) is ~1% of all resistant colonies (ie colonies that have taken up the DNA seq used to make the knock in), when lipofectamine reagents are use. Lipofectamin trasfromation efficiency varies depending on the size of the transformed DNA, host cell type and plating density. So for a 5kb construct that transformation efficiency varies between 30%-50%.

CRISPR is nearly 100% efficient in yeast cells, ie all knock in cells are correctly targeted. This is likely due to the high levels of HR in yeast