This is a great breakthrough for modern science.
American researchers have developed a computational model that predicts the outcome of the reparation of a particular DNA sequence after it was cut a protein called Cas9.
As it turned out, for about a tenth of cases with a high probability to predict what sequence there is formed after the operation of systems of DNA reparation. This allowed the scientists to fix the number of harmful mutations in human cells with CRISPR without the use of a matrix to edit.
System for genome editing CRISPR-Cas9 has two main components, a protein called Cas9 and a short seed (the guide RNA), which specifies the Cas9 which point he cut. This basic set is, strictly speaking, nothing to edit, it just brings in the genome of double-stranded break at a specified location. To insert in this place the right order is necessary for the third component matrix of DNA containing the sequence to be gene. Using this matrix, the cell repair system by the mechanism of homologous recombination heals the gap in the DNA and builds the desired piece.
In the absence of a matrix for reparations (and even if it is as homologous recombination in human cells works pretty bad) cut is restored with the participation of other DNA repair systems, in particular system connections homologous end joining (NHEJ) and the connection end on the basis of microhomology (MMEJ). After operation of these systems at the incision remains small deletions or insertions, which in most cases disrupt the gene. That is why with the help of “the core set of” CRISPR-Cas9 is easy to break a gene, but hard to repair.
Researchers from the Massachusetts Institute of technology decided to turn the disadvantage of the latter in dignity and created a model based on machine learning, which with high probability predicts the outcome of DNA repair mechanisms NHEJ and MMEJ, that is, according to what kind of sequence at the site of incision is formed after reparations taking into account deletions and insertions of at least 50 percent of the time. According to the model, to predict the outcome of reparations with such precision is possible for 5-11% of all guide RNAS to the human genome (“precise-50”). To build the model inDelphi the researchers used experimental data that was obtained after the cutting of the Cas9 genome nearly two thousand sites in the DNA.
After you create the model, scientists experimentally confirmed its relevancy – it from the list of “precise-50”-guide RNA selected 14 that “incited” to Cas9 for sequence mutation (namely, microdeletions in one nucleotide), characteristic for a particular genetic disease. After repair of rupture in this place, according to inDelphi should receive the extra nucleotide. It turned out that indeed, after work and CRISPR mismatch repair system gene sequence was recovered due to this microinserts on average in 60 percent of cases.
This means that some harmful mutations (deletions or insertions), leading to the development of disease can be corrected with CRISPR without the use of a matrix for reparations and with a sufficiently high efficiency. Only the researchers was able to find the guide RNA from the list of “precise-50” 195 such harmful alleles and experimentally confirmed that they are with a frequency of more than 50 percent corrected to normal after the cut and repair. For example, they managed to edit a mutation in the HPS1 gene in fibroblasts of patients with the syndrome of the German-Pudlak, leading to skin pigmentation disorders and hemophilia, and mutations in the gene ATP7A in cells of patients with Menkes ‘ disease.
To edit the genome without the use of the matrix is also using so-called “editors of reason” on the basis of CRISPR-Cas, which are able to fix all types of nucleotide substitutions. We wrote, for example, with the help of this tool adult mice cured from phenylketonuria.