Anshul Kundaje argues trained models recover causal sequence syntax from in vivo data
anshulkundaje · x · 2026-11-19
In a scholarly exchange, a critic argues it is extremely difficult, if not impossible, to convincingly infer causal determinants of transcription factor binding from in vivo observations of native DNA, insisting on in vitro experiments with controlled synthetic DNA mixtures and TF mutants. Stanford computational biologist Anshul Kundaje concedes there are no causality guarantees but points to several published studies at non-trivial scale showing recovery of causal sequence syntax from carefully trained models of observational in vivo data, validated with in vitro and in-genome perturbation experiments. A representative debate on causal inference limits of AI models in genomics.
More from Research
- François Fleuret: perfect inductive bias causes brutal loss drops that gradient descent reads as spikes — francoisfleuret · 2026-11-19
- WebGPU playground lets you experiment with weight-space programming on Qwen3.5-0.8B — wsxiaoys · 2026-11-19
- ViT can express any conv layer mathematically, but CNN priors still win on small edge devices — burny_tech · 2026-11-19
- Eval Cards tracks 309K model-benchmark results; 78% of evals can't be reproduced — davidmanheim · 2026-11-19
- Jason Wei maps which benchmarks frontier labs actually report; four see universal adoption — davidmanheim · 2026-11-19
- PIPETTE fixes RL red-teaming mode collapse to harden LLM agents against prompt injection — Lipeng He · 2026-11-19