Downloads · 30 days
36
11% of all-time downloads
ibm-research/biomed.dna.snp.modernbert.113m.v1
biomed.dna.snp.modernbert.113m.v1 is a machine learning model from ibm-research. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for biomed-multi-omic. The card lists the license as apache-2.0.
Biomedical foundational models for omics data. This package supports the development of foundation models for scRNA or for DNA data.
Downloads · 30 days
36
11% of all-time downloads
All-time downloads
334
Public
Repo size
1.4 GB
Likes
5
Public
Click a slice to open those files.
.ckpt1.4 GB · 100%
From the Hugging Face model README
Biomedical foundational models for omics data. This package supports the development of foundation models for scRNA or for DNA data.
biomed-multi-omic enables development and testing of foundation models for DNA sequences and for RNA expression,
with modular model and training methods for pretraining and fine-tuning, controllable via a declarative no-code interface.
biomed-multi-omic leverages anndata, HuggingFace Transformers, PyTorchLighting and Hydra.
For details on how the models were trained, please refer to the BMFM-DNA preprint.
BMFM-DNA-SNP
We extracted 20 million genetic variants from Single Nucleotide Polymorphism Database (dbSNP), including SNPs, insertions and deletions. For each genomic position that has variants, we averaged available population-specific variation frequencies and built a genome-wide variation frequency matrix. For each genomic position, this matrix provides the multinomial distribution of 11 probabilities on 5 types of nucleotides (A, C, G, T, N), 5 types of insertions (insertions after A, C, G, T, N), and deletion. For genomic positions that do not have alternatives, we assign the probability of 1.0 to the nucleotide based on the human reference genome GRCh38. Through this genome-wide variation frequency matrix, we can efficiently retrieve biallelic variants based on population frequencies and sample variant-encoded DNA sequences.
For full details see section 3.1.2 of the BMFM-DNA manuscript.
Using biomed.dna.snp.modernbert.113m.v1 requires the codebase https://github.com/BiomedSciAI/biomed-multi-omic.
For installation, please follow the instructions on github.
To get embeddings for DNA sequences run:
export INPUT_DIRECTORY=... # path to your DNA sequences files
bmfm-targets-run -cn dna_predict input_directory=$INPUT_DIRECTORY working_dir=/tmp checkpoint=ibm-research/biomed.dna.snp.modernbert.113m.v1
For more details see the DNA tutorials on github.
To cite the tool for both RNA and DNA, please cite both the following articles:
@misc{li2025bmfmdnasnpawarednafoundation,
title={BMFM-DNA: A SNP-aware DNA foundation model to capture variant effects},
author={Hongyang Li and Sanjoy Dey and Bum Chul Kwon and Michael Danziger and Michal Rosen-Tzvi and Jianying Hu and James Kozloski and Ching-Huei Tsou and Bharath Dandala and Pablo Meyer},
year={2025},
eprint={2507.05265},
archivePrefix={arXiv},
primaryClass={q-bio.GN},
url={https://arxiv.org/abs/2507.05265},
}