Find open-source science resources
A directory of tools, AI models, datasets, and research resources for biotech, bioinformatics, and other scientific fields. Aggregated from curated GitHub awesome-lists, HuggingFace, bio.tools, Bioconductor, and more.
Filters
Health
Domain
Language
License
Source
Type(1)
1,095 of 9,874 resources
Showing 1–50
genome-spy/BPNet-ONNX
by genome-spyConversion finished: 11,695 accepted / 11,695 source checkpoints. All 2,339 experiments were attempted. The release has 0 exclusions, recorded in excluded-checkpoints.json. accepted-manifest.json lists the exact accepted artifact identities, hashes, validation errors and converter provenance.
genome-spy/ChromBPNet-ONNX
by genome-spyConversion finished: 7,560 accepted / 7,560 source checkpoints. All 1,512 experiments were attempted. The release has 0 exclusions, recorded in excluded-checkpoints.json. accepted-manifest.json lists the exact accepted artifact identities, hashes, validation errors and converter provenance.
In recent years, pre-trained language models (PLMs) achieve the best performance on a wide range of natural language processing (NLP) tasks. While the first models were trained on general domain data, specialized ones have emerged to more effectively treat specific domains.
genome-spy/BATv1-ONNX
by genome-spyONNX conversions of the published BATv1 atlas for predicting human transcription factor / DNA-associated protein binding (ChIP-seq). Original models and metadata are by programmable-genomics / the Cherimoya authors. GenomeSpy provides serialization and numerical validation; this is not an…
InstaDeepAI/MOJO
by InstaDeepAIMOJO (MultiOmics JOint representation learning) is a model that learns joint representations of bulk RNA-seq and DNA methylation through bimodal masked language modeling and is tailored for cancer-type classification and survival analysis on the TCGA dataset.
OpenMed GLiNER2.5 PII Multilingual v1
OpenMed GTE-ModernBERT Biomedical Embedding Base v1
OpenMed EmbeddingGemma2 BiomedicalEvidence 270M v2
OpenMed GLiNER2.5 ProceduresTests Small v1
xinHesysu/bioseq-models
by xinHesysuPretrained weights consumed by the bioseq-models Python package. The package downloads only the files required by the model selected by the user.
liuyong9975/Calpha
by liuyong9975Toward native biomolecular intelligence in a general-purpose language model. Technical Report I: Protein Folding.
beapen/gt-embed-small-v1
by beapenSmall CDS embedding model balanced for performance and efficiency. Trained using contrastive learning with the Genomic Tokenizer on human protein-coding DNA sequences.
Accurately transcribes English doctor–patient conversations. It recognises medical conditions correctly 94.6% of the time (base model: 83.9%) and medicine names 66.1% (base: 61.2%) in consultations, while matching the base model's accuracy on everyday words.
> Requires the Google Gemini API. These SAEs take gemini-embedding-001 embeddings as input. That model is proprietary and only available through the paid Gemini API or Vertex AI; no weights are distributed with this repository.
Sparse autoencoders (SAEs) trained on abhinand/MedEmbed-large-v0.1 embeddings for medical retrieval, from the paper Explanations that Transfer: Interpreting Medical Retrieval with Sparse AutoEncoders (ML4H 2026). Code: doctolib-lab/doctosae.
Sparse autoencoders (SAEs) trained on BAAI/bge-large-en-v1.5 embeddings for medical retrieval, from the paper Explanations that Transfer: Interpreting Medical Retrieval with Sparse AutoEncoders (ML4H 2026). Code: doctolib-lab/doctosae.
allforcloud9527/ocsu-hybrid-fg-convnext-t
by allforcloud9527Multi-label image classifier for the 140 functional-group names used by the OCSU paper. It replaces the OCSR -> RDKit dictionary stage with an image-only head, so functional groups do not depend on SMILES quality.
CRG-MIT/ICICLE
by CRG-MITnvidia/RFD4-Proteina
by nvidiaSequence and structure, designed together.Baker Lab × NVIDIA GenAIR Project page · Code · Documentation · Training data · Studio (ZeroGPU) · Studio (dedicated GPU)
Eyened/vascx
by Eyened> [!IMPORTANT] > The entire VascX pipeline, including feature extraction is available here: retinalysis-vascx. A new preprint presenting the pipeline is available here. We will stop supporting this repository and move to the new one linked above. We will keep updating the models and pipeline there.
OpenMed GLiNER2.5 Anatomy Small v1
zatumout/pooper-b10-1b
by zatumoutA ~983M-parameter, 37-layer-tower model that predicts 1,885 ENCODE signal tracks (ATAC-seq, DNase-seq, histone ChIP-seq, RNA-seq; GRCh38) at 1 bp resolution from DNA sequence, using an AlphaGenome-style architecture (arch = alphagenome).
mradermacher/Qwen3.5-9B-DeepThought-STEM-i1-GGUF
by mradermacherFor a convenient overview and download list, visit our model page for this model.
mradermacher/Qwen3.5-9B-DeepThought-STEM-GGUF
by mradermacherFor a convenient overview and download list, visit our model page for this model.
mradermacher/Qwen3.5-4B-DeepThought-STEM-i1-GGUF
by mradermacherFor a convenient overview and download list, visit our model page for this model.
mradermacher/Qwen3.5-4B-DeepThought-STEM-GGUF
by mradermacherFor a convenient overview and download list, visit our model page for this model.
ashford-andrew/UniVI-PBMC-Multiome
by ashford-andrewInteractive explorer · Source · Paper
OpenMed GLiNER2.5 DrugChemical Small v1
arcinstitute/ST-SE-Replogle
by arcinstituteState Transition (ST) models using X_state inputs on Replogle. This repository contains 4 fewshot and 4 zeroshot canonical models, with their saved configurations, loading files and available CellEval v1 outputs.
arcinstitute/ST-SE-Parse
by arcinstituteState Transition (ST) models using X_state inputs on Parse. This repository contains 4 fewshot and 4 zeroshot canonical models, with their saved configurations, loading files and available CellEval v1 outputs.
Godseye1311/UniHyena-checkpoints
by Godseye1311ChatterjeeLab/CupOFMoCA
by ChatterjeeLabCode, model weights, configurations, and reported outputs for CupOFMoCA: Coupled Objective-Guided Discrete Flows for Molecular Conjugate Assembly (NeurIPS 2026).
CladeTeam/CENO-9B-A3B-1m
by CladeTeamCENO-9B-A3B-1m is the long-context (1M) checkpoint of the 9B CENO DNA foundation model — a causal language model over genomic sequence built on a Nemotron-H Mamba / Attention / Mixture-of-Experts hybrid backbone (no MSA inputs).
CladeTeam/CENO-9B-A3B-131k
by CladeTeamCENO-9B-A3B-131k is the long-context (131k) checkpoint of the 9B CENO DNA foundation model — a causal language model over genomic sequence built on a Nemotron-H Mamba / Attention / Mixture-of-Experts hybrid backbone (no MSA inputs).
CladeTeam/CENO-9B-A3B-base
by CladeTeamCENO-9B-A3B-base is the base pretraining checkpoint of the 9B CENO DNA foundation model — a causal language model over genomic sequence built on a Nemotron-H Mamba / Attention / Mixture-of-Experts hybrid backbone (no MSA inputs).
CladeTeam/CENO-1.3B-A450M-1m
by CladeTeamCENO-1.3B-A450M-1m is a checkpoint of the CENO base DNA foundation model (1M context (stage 4)). It is a plain causal language model over genomic sequence on a Nemotron-H Mamba/Attention/MoE hybrid backbone, with no MSA inputs.
CENO-1.3B-A450M-131k is the long-context (131k) checkpoint of the 1.3B CENO DNA foundation model — a causal language model over genomic sequence built on a Nemotron-H Mamba / Attention / Mixture-of-Experts hybrid backbone (no MSA inputs).
CENO-1.3B-A450M-base is the base pretraining checkpoint of the 1.3B CENO DNA foundation model — a causal language model over genomic sequence built on a Nemotron-H Mamba / Attention / Mixture-of-Experts hybrid backbone (no MSA inputs).
CladeTeam/CENO-600M-A230M-1m
by CladeTeamCENO-600M-A230M-1m is the long-context (1M) checkpoint of the 600M CENO DNA foundation model — a causal language model over genomic sequence built on a Nemotron-H Mamba / Attention / Mixture-of-Experts hybrid backbone (no MSA inputs).
CENO-600M-A230M-131k is the long-context (131k) checkpoint of the 600M CENO DNA foundation model — a causal language model over genomic sequence built on a Nemotron-H Mamba / Attention / Mixture-of-Experts hybrid backbone (no MSA inputs).
CENO-600M-A230M-base is the base pretraining checkpoint of the 600M CENO DNA foundation model — a causal language model over genomic sequence built on a Nemotron-H Mamba / Attention / Mixture-of-Experts hybrid backbone (no MSA inputs).
CladeTeam/CENO-300M-A105M-1m
by CladeTeamCENO-300M-A105M-1m is the long-context (1M) checkpoint of the 300M CENO DNA foundation model — a causal language model over genomic sequence built on a Nemotron-H Mamba / Attention / Mixture-of-Experts hybrid backbone (no MSA inputs).
CENO-300M-A105M-131k is the long-context (131k) checkpoint of the 300M CENO DNA foundation model — a causal language model over genomic sequence built on a Nemotron-H Mamba / Attention / Mixture-of-Experts hybrid backbone (no MSA inputs).