Single-cell driven enhancer modelling and design

The combination of scRNA-seq and scATAC-seq allows building gene-regulatory atlases of any tissue and species. I will present several new computational strategies that exploit single-cell multi-omics data: (1) to model genomic enhancers using topic modelling and convolutional neural networks; and (2) to derive “enhancer-GRNs” (eGRN) with key transcription factors, genomic enhancers, and predicted target genes per cell type. I will discuss the results of several case studies where we applied these strategies, including the Drosophila brain, human melanoma, the mouse liver, and the evolution of cell types in the vertebrate telencephalon. Finally, I will discuss how enhancer models based on deep learning can be exploited to design synthetic enhancers for Drosophila and human cell types.

---

From 30 Apr 2024 11:00
Until 30 Apr 2024 12:30
Location FSVM I building, seminar room

Speaker Stein Aerts
Affiliation VIB Center for AI & Computational Biology, KU Leuven
Host Dirk Inzé & Yves Van de Peer