jaxQTL in Nature Genetics: count-based models find 16% more eGenes in single-cell eQTL
anshulkundaje · x · 2026-11-11
The Kundaje lab's jaxQTL paper is out in Nature Genetics. The framework uses highly efficient JAX-based count models (negative binomial) for pseudobulk single-cell eQTL mapping, addressing the problem that existing approaches are either not designed for sparse single-cell RNA-seq counts or become computationally intractable at scale.
In simulations, jaxQTL's negative binomial model provides calibrated inference while outperforming linear models. On the OneK1K dataset, it identifies 16% more eGenes than linear models while running substantially faster, and reveals cell-type-specific genetic regulation.
More from Research
- TRIAD benchmark: no multimodal retrieval system leads across its six subsets — _reachsumit · 2026-11-11
- Vision tokens far denser than text: paper traces VLM modality gap to early-layer processing — _reachsumit · 2026-11-11
- ByteDance's SepLRM splits recommendation models around a shared KV cache — _reachsumit · 2026-11-11
- Yuanbao-Reranker: 0.6B/4B/8B Qwen3-based rerankers trained on synthetic preferences, open-sourced — _reachsumit · 2026-11-11
- Shopify's RoutingTrie lifts generative retrieval Hit@100 from 67.5% to 96.1% with 1.1ms overhead — _reachsumit · 2026-11-11
- Hybrid product search straight from the data lake: Pinot + BM25 + RRF hits 6.5-9ms p50 on 277M products — _reachsumit · 2026-11-11