AI improved editor of genome, predicting, rezultatas editor of the genome of the CRISPR-Cas9 is pretty well learned to deal. However, with a fairly good knowledge of the technology and the principles of its work, the result that will be obtained in the course of editing is not always predictable. This is due to the peculiarities of the editing process and its outcome learned to predict artificial intelligence, further improving the effectiveness of the method of editing the genome.
First, I would like to explain what the main problem is. Oddly enough, the editor of the genome there is almost nothing to do with it. Cas9 protein in this case serves as a blade and cuts out the required component of the genome. But then on the” empty space ” you need to insert another component. It uses matrix DNA, which, if explained in simple language, acts as a” donor ” materials. This process is called DNA repair. There are other systems that work on a similar principle in the absence of matrix DNA, but it is in them that the problem lies: after their work on the site of replacement of genetic material, deletions may remain (roughly speaking, missing areas). The outcome of such interventions previously could only be calculated empirically.
According to the editorial Board of the journal Nature, a group of scientists from the Massachusetts Institute of Technology (MIT) has developed a program based on artificial intelligence, which calculates the result of genome changes. AI is able to report what sequence is formed in the place of intervention after editing, as well as in at least 50% of cases it reports whether deletions can remain after the operation is completed. In addition, if you use AI to predict all stages, the prediction accuracy can be increased to 5-11%, and this is a very good result.
In order to teach AI to predict the outcome of the intervention in the genome, scientists have created a library of 2000 RNA guides for Cas9 (RNA indicates the protein, where it is necessary to cut the molecule). Next, 14 of them were selected, which could cause the appearance of excess nucleotide after repair, and then the AI chose those that most often led to repair errors. Based on this data, a “behavior model” was built, which was already used for other operations. Scientists have managed effectively and without any side effects edit of mutation in cells with the syndrome of the German-Pudlak and disease Menekse.