There is a newer version of the record available.

Published April 22, 2026 | Version v2
Dataset Open

D3-DNA: Data and Model Weights for DNA Discrete Diffusion

Authors/Creators

  • 1. Cold Spring Harbor Laboratory

Description

Data files and pretrained model weights for D3: DNA Discrete Diffusion, a discrete diffusion framework for conditional generation of DNA regulatory sequences.

Code: https://github.com/anirbansarkar-cs/D3-DNA-Discrete-Diffusion
Package: d3-dna (pip install d3-dna)

This deposit contains LentiMPRA datasets (K562, HepG2, WTC11), LegNet oracle checkpoints for evaluation, a DeepSTARR oracle, and the D3-Transformer diffusion model trained on K562. Additional datasets and model weights will be added in future versions.

Files

README.md

Files (16.6 GB)

Name Size Download all
md5:23bd19ea598805d0ac1ef2cd3d414f2c
214.1 MB Download
md5:c4c74a36daa44e63b4a616568e1995b4
201.9 MB Download
md5:fb883c4f3114dc6b4757c546d9642a44
1.5 GB Download
md5:f8f13d2f8145478fe322587eed5d1e15
1.5 GB Download
md5:e282c76933471a005f4ab821053ce4ee
1.5 GB Download
md5:de40be87592cc77cb8ecf4e269dabac9
1.5 GB Download
md5:f9d0cae9ce4dd3c90b465e407d842cc7
294.6 MB Download
md5:b37762d49dac75e4036c02398b468c8b
970.7 MB Download
md5:da0df7173e2e00aad4e37c8995af7ee4
1.6 GB Download
md5:ebd9ff5a884a30cc63ae645856063e76
246.0 MB Download
md5:d9b50c4f358e8bef72e1eff8d3951630
7.5 MB Download
md5:37ca20c28d8f99ac72401f784bf0869c
16.0 MB Download
md5:1b7b52d9416bb88d4eb8f5740b8d1d6f
16.0 MB Download
md5:e43843b40c904956f39fc6edd48055ac
7.1 GB Download
md5:1e04b35254138b1e9ea30dd794fd3803
6.2 kB Preview Download

Additional details