Fabian Theis says Virtual Cell work should target harder problems than perturbation prediction
Overview
Fabian Theis says perturbation prediction on existing datasets is not a hard enough problem to justify complex algorithms, according to a Latent Space report dated November 6, 2026.
He argues the Virtual Cell field should instead aim to predict how molecules, cells, niches, tissues, organs and organisms affect one another.
Theis links this goal to the $1.8B Biohub data collection effort, which he says is designed to support such predictions. The report presents his position as his own view and does not include independent evaluation of the argument.
Written by AI from the articles below · updated Nov 6, 10:29 AM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- Latent SpaceBlogFabian Theis argues Virtual Cell work needs harder problems than perturbation prediction
AIFabian Theis believes perturbation prediction on existing datasets is not a hard enough problem to justify complex algorithms. He says the field should instead target predicting how molecules, cells, niches, tissues, organs and organisms affect one another, which the $1.8B Biohub data collection effort is designed to support.
Heat Trend
Not enough continuous observations to show a trend yet.