Modeling and design of genomic enhancers in the context of whole organisms

Datum

04.09.2025

Zeit

11:00 - 12:00

Sprecher

Stein Aerts

Zugehörigkeit

VIB Center for Computational Biology and AI, Belgium

Serie

MPI-CBG Thursday Seminar

Sprache

en

Hauptthema

Biologie

Host

Anna Poetsch/Merixtell Huch

Beschreibung

I will present several new computational strategies to train deep learning models using single-cell multi-omics training data. I will discuss new software packages, CREsted and TFMInDI, to train enhancer models, design synthetic enhancers, and decipher cis-regulatory logic. I will illustrate these approaches on diverse biological systems, including the mammalian cortex, zebrafish development, and early human neural development. We further explore how enhancer models can be used to study the evolution of cell types. Next, we look at enhancer models in the context of human personal genomes, to prioritize and understand cis-regulatory variation by combining enhancer models with whole genome sequencing and QTL analysis. Finally, we investigate enhancer AI models in the context of foundation models to also predict (single-cell) gene expression, applied to the Fly Cell Atlas.

Letztmalig verändert: 25.08.2025, 07:40:06