Deep learning suggests that gene expression is encoded in all parts of a co-evolving interacting gene regulatory structure
Problem
It was unclear how coding and non-coding regions of the gene regulatory structure interact to determine mRNA expression levels.
Why it matters
Decoding gene regulation is fundamental to understanding molecular biology and disease.
What the AI did
Deep learning was applied to over 20,000 mRNA datasets to predict mRNA abundance directly from DNA sequence across 7 model organisms, and to search for regulatory motif interactions.
Results
Up to 82% of transcript-level variation predicted from DNA sequence; motif interactions could explain the whole dynamic range of mRNA levels.
Impact
Provides evidence that the entire co-evolving gene regulatory structure, not single motifs, defines gene expression.
Record
| Venue | Nature Communications |
|---|---|
| Published | 2020 |
| Work type | article |
| Authors | Jan Zrimec, Christoph S Börlin, Filip Buric, Azam Sheikh Muhammad, Rhongzen Chen, Verena Siewers, Vilhelm Verendel, JENS KVIST NIELSEN et al. |
| Citations | 172 (OpenAlex, at import time) |
| References | 158 |
| Model or method | not stated in source |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41467-020-19921-4 |
| Open access | gold |
| Keywords | deep learning, gene regulation, mRNA abundance, regulatory motifs, genomics |
Importance rating
significant — thousands to millions affected
Abstract frames the genetic regulatory code as an important molecular biology challenge; analysis of 20,000+ mRNA datasets across 7 organisms and 82% of variance explained indicate broad foundational relevance.