MIT team uses sequence-to-function deep learning to decode human cis-regulatory evolution

burny_tech · x · 2026-09-28

Researchers from MIT CSAIL, the Broad Institute and Harvard (including Manolis Kellis) posted a bioRxiv preprint applying sequence-to-function deep learning to decode human cis-regulatory evolution. Led by Riley J. Mangan et al., the work uses deep models on regulatory DNA sequences to study how human cis-regulatory regions have evolved — an AI-for-Science study in genomics.

Original post →

More from Research

Research channel →