@alexanderfuxi @anshulkundaje @jeffvierstra Contact frequency of random pieces of DNA strongly depends on 1D distance but specific sites positioned at long distance (10kb-1mb) can be high contact frequency. Enhancer CRISPRi suggests a relatively weak dista
4/ This training objective could address some challenges in the field raised by @AlexKarollus & @LXandR_ in their stellar works. I'd love to see the impacts of scaling the model in terms of parameters and RNA-seq sample count! ⚖️ https://t.co/uHOoKniu
RT @gagneurlab: Deep models as in silico theories of regulation? We checked how Enformer, trained on the human genome, holds up to experime…
RT @gagneurlab: Deep models as in silico theories of regulation? We checked how Enformer, trained on the human genome, holds up to experime…