AlphaGenome is fine-tuned on RNA-seq modalities beyond coverage tracks
anshulkundaje · x · 2026-07-29
A genomicsxai blog post from Miquel Anglada-Girotto describes a framework for preprocessing, loading, and fine-tuning AlphaGenome on multiple RNA-seq modalities, including gene expression, splice-site usage, and splice junctions.
The post explains the motivation: adapting AlphaGenome to RNA-seq samples that were not part of pretraining, and evaluating whether the model can learn sample-specific multimodal transcriptomic patterns on held-out genomic intervals. As a case study, it uses an SF3B1 K700E cancer-driver sample and a wild-type control, because the mutation is known to alter splicing and promote cryptic splice sites.
Related event: New Fine-Tuning Framework Adapts AlphaGenome to Diverse RNA-seq Modalities(2 posts)→
More from Research
- LLMs on robots jump real-world success from 16.7% to 97.3% — tri_dao · 2026-07-29
- Simple teleoperation recordings plus soft compliance can solve more robot tasks than expected — ihorbeaver · 2026-07-29
- Studies find AI therapy replies often score higher on empathy than human clinicians — sapinker · 2026-07-29
- TransluceAI proposes oversight foundation models to catch reward hacking at scale — JacobSteinhardt · 2026-07-29
- Student builds a local coding agent on molab and beats two 7B coder baselines — S_Conradi · 2026-07-29
- Why tokenizer optimization is hard: expensive pretraining, slow feedback, and non-differentiable design — paul_cal · 2026-07-29