Machine Learning Discovers Sequences to Boost Drug Delivery

Researchers at the Massachusetts Institute of Technology (MIT) have developed an approach that uses experimental chemistry and artificial intelligence to facilitate delivery of a drug to treat Duchenne muscular dystrophy (DMD).

The approach aims to determine which cell-penetrating peptides can be attached to antisense phosphorodiamidate morpholino oligomers (PMO) to help permeate the cell nucleus to modify a mutated dystrophin gene and enable production of a key protein typically missing in patients with DMD.

The researchers built a library of 600 miniproteins, each attached to PMO, by mixing and matching 57 peptides, and quantified each miniprotein's ability to deliver drugs across the cell.

MIT's Rafael Gomez-Bombarelli said, "The key innovation is using machine learning to connect the sequence of a peptide, particularly a peptide that includes non-natural amino acids, to experimentally-measured biological activity."

The sequences proposed by the model were found to be more effective than previously known variants.

From MIT News
View Full Article

Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA