A bigwig-style file format built for genomic ML with direct GPU loading
jmschreiber91 · x · 2026-10-28
A thread describes a new file format built for modern genomic machine learning: it can do anything a bigwig can (zooms, arbitrary data types, sparsity along either of the first two dimensions, pseudobulk extraction for single-cell/molecule data), and enables GPU loading via a general-purpose kernel compatible with JAX, TensorFlow and PyTorch, with multiple loading paths including pinned memory and dlpack. The author admits the current benefit is modest.
Related event: New File Format Brings GPU-Native Loading to Genomic ML(2 posts)→
More from Research
- NVIDIA open-sources Flora, a model suite that auto-labels robot videos for training GR00T — qinzytech · 2026-10-28
- Multimodal GraphRAG models missing and contradictory evidence for storm damage assessment — JeremyCMorgan · 2026-10-28
- SIMIT: VLMs self-improve at test time by imagining practice problems, +7.2% on 17 benchmarks — delliott · 2026-10-28
- Mistral sponsors Open Math OS with zero data retention, powering a Kimi K3/DeepSeek math platform — GuillaumeLample · 2026-10-28
- CROP wins EMNLP 2026 Outstanding Paper: conformal certification for reasoning trace prefixes — hanjie_chen · 2026-10-28
- OSU NLP Group Wins Three Paper Awards Same Day at EMNLP 2026 and NeurIPS Workshop — hhsun1 · 2026-10-28