Back

Applications of machine learning to solve genetics problems

Sowunmi, K.; Soyebo, T. A.; Okosesi, E. A.; Adesiyan, A. L.; Oladimeji, K. A.; Ajibola, O. A.; Ogunlana, Y. O.; Agboola, O. W.; Kaur, G.; Atoromola, H.; Oladipupo, T. A.

2020-10-27 bioinformatics
10.1101/2020.10.27.354092 bioRxiv
Show abstract

The development of precise DNA editing nucleases that induce double-strand breaks (DSBs) - including zinc finger nucleases, TALENs, and CRISPR/Cas systems - has revolutionized gene editing and genome engineering. Endogenous DNA DSB repair mechanisms are often leveraged to enhance editing efficiency and precision. While the non-homologous end joining (NHEJ) and homologous recombination (HR) DNA DSB repair pathways have already been the topic of an excellent deal of investigation, an alternate pathway, microhomology-mediated end joining (MMEJ), remains relatively unexplored. However, the MMEJ pathways ability to supply reproducible and efficient deletions within the course of repair makes it a perfect pathway to be used in gene knockouts. (Microhomology Evoked Deletion Judication EluciDation) may be a random forest machine learning-based method for predicting the extent to which the location of a targeted DNA DSB are going to be repaired using the MMEJ repair pathway. On an independent test set of 24 HeLa cell DSB sites, MEDJED achieved a Pearson coefficient of correlation (PCC) of 81.36%, Mean Absolute Error (MAE) of 10.96%, and Root Mean Square Error (RMSE) of13.09%. This performance demonstrates MEDJEDs value as a tool for researchers who wish to leverage MMEJ to supply efficient and precise gene knock outs.

Matching journals

The top 6 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.