CRISPR meets machine learning

If a donor template is not provided when CRISPR cuts the DNA, broken ends are fixed by natural repairing mechanisms in a way that is considered stochastic and heterogeneous. This makes template-free editing impractical beyond gene disruption, right? Wrong, according to a study published in Nature by Richard Sherwood and colleagues.

The researchers used a library of nearly 2.000 paired Cas9 guide RNAs and human DNA target sites to train a machine learning model called inDelphi. It predicts that 5-11% of guides targeting the human genome can induce a single repair outcome in over 50% of instances. The same model could be used to identify suitable pathogenic mutations to fix using template-free editing. Scientists proved their case by accurately repairing nearly 200 pathogenic variants related to three diseases: Hermansky-Pudlak syndrome, Merkes disease and familial hypercholesterolemia.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s