香蕉视频

Learning single cell atlases

Fabian J. Theis, Helmholtz Munich
Tuesday October 26, 12-1pm
Zoom Link:聽

Abstract:聽Advances in single cell genomics nowadays allow the large-scale construction of organ atlases. These can be used to study perturbations such as signaling, drugs or diseases, with large-scale access to state changes on the transcriptomic and more recently also epigenomic level. This provides an ideal application area for machine learning methods to understand cellular response. Here I will first show how we leverage representation learning to identify a gene expression manifold, and thus build a lung cell atlas across datasets from many labs. I will demonstrate how to query this atlas for new biology using transfer learning, and then discuss how to include other modalities, spatial context as well as perturbations into such models.