But is regulatory grammar conserved? We transfected the exact same libraries into matched primary human cells. We found widespread conservation of autonomous regulatory activity & allelic effects across species. Suggesting human data/models can be leveraged to inform cattle.
Supporting this, we used this massive dataset to train deep-learning models predicting promoter activity from sequence. Models trained on data from cattle and human cells learnt broadly similar regulatory grammar.
Huge thanks to the brilliant team at the @roslininstitute.bsky.social, and our collaborators! This resource is a major step for prioritizing causal variants in livestock. Check out the full preprint: www.biorxiv.org/content/10.6...