Bharath Ramsundar: Today's 'Virtual Cells' Are Nowhere Near Physically Realistic
rbhar90 · x · 2026-09-18
Bharath Ramsundar, founder of Deep Forest Sciences, publishes a new essay examining the virtual cell hype. He shares the excitement but argues that what is now called a "virtual cell" is not really a mechanistic model, and we remain far from physically realistic cell simulation.
- He recalls Markus Covert's pioneering Stanford whole-cell in silico simulation: the GitHub repo was a tangle of Matlab scripts with constants pulled from literature or painstakingly estimated; the lab continues with an actively developed E. Coli whole-cell model
- Current transformer-based efforts modeling perturbation networks in cell lines are interesting research projects, but fall short of capturing actual cellular dynamics
- He argues that broadly generalizable, physically grounded virtual cells will be crucial for ambitious drug discovery efforts aiming at fundamental progress in the coming years
A sobering AI-for-Science reality check for anyone working on computational biology.
More from Research
- PixelBench: testing multimodal models pixel-by-pixel instead of with QA — bdsqlsz · 2026-11-04
- François Chollet: modeling brain wiring as an engineering problem constrained by genome size and time — CatAstro_Piyush · 2026-11-04
- OffsecCyber dataset combines HackerOne reports and CVEs to train pentesting LLMs — AccBalanced · 2026-11-04
- Monofact rate in training data governs LLM scientific discovery, preprint finds — CatAstro_Piyush · 2026-11-04
- qwenTPU compiles a TPU into Qwen3.5 weights, achieving exact integer matmul — CatAstro_Piyush · 2026-11-04
- Jev model improves ICL demonstration retrieval via distillation into a lightweight MLP — kalyan_kpl · 2026-11-04