BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Estimation of Single-cell and Tissue Perturbation Effect in Spatia
 l Transcriptomics via Spatial Causal Disentanglement - Stathis Megas\, Wel
 lcome Sanger Institute
DTSTART:20241018T120000Z
DTEND:20241018T130000Z
UID:TALK222568@talks.cam.ac.uk
CONTACT:Ferdia Sherry
DESCRIPTION:Models of virtual cells and virtual tissues at single-cell res
 olution would allow us to test perturbations in silico and accelerate prog
 ress in tissue and cell engineering. However\, most such models are not ro
 oted in causal inference and as a result\, could mistake correlation for c
 ausation. We introduce Celcomen\, a novel generative graph neural network 
 grounded in mathematical causality to disentangle intra- and inter-cellula
 r gene regulation in spatial transcriptomics and single-cell data. Celcome
 n can also be prompted by perturbations to generate spatial counterfactual
 s\, thus offering insights into experimentally inaccessible states\, with 
 potential applications in human health. We validate the model’s disentan
 glement and identifiability through simulations\, and demonstrate its coun
 terfactual predictions in clinically relevant settings\, including human g
 lioblastoma and fetal spleen\, recovering inflammation-related gene progra
 ms post immune system perturbation. Moreover\, it supports mechanistic int
 erpretability\, as its parameters can be reverse-engineered from observed 
 behavior\, making it an accessible model for understanding both neural net
 works and complex biological systems.
LOCATION:MR2 Centre for Mathematical Sciences
END:VEVENT
END:VCALENDAR