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
9,874 resources indexed
Showing 9,851–9,874
Web-based and desktop molecular biology software for plasmid design, molecular cloning, sequence alignment, genome browsing, CRISPR-Cas guide RNA design, PCR primer design, virtual gel electrophoresis simulation, and reverse translation from protein to DNA. The application includes the ability to create teams and work in a shared workspace.
Python toolkit for processing , analyising and visualising high-speed AFM (Atomic Force Microscopy) video data.
Browser-based workspace for sequence analysis and format converters for researchers and students. Homology search, multi-FASTA gene trees, motif and domain correlation, Golden Gate quality control, and sequence format conversion, with a focus on plant pathology and comparative genomics. Guest access — try without signing in.
Detects swapped tubes, mislabelled samples and sample-sheet shifts in longitudinal microbiome studies (16S and shotgun) from each subject's microbial fingerprint. It self-calibrates by planting errors in the user's own labels, reports its own detection power, and runs entirely in the web browser with no data uploaded. An R package (swapseqr) runs the same engine on phyloseq objects or count matrices.
spaCR links pooled CRISPR-Cas9 perturbations to single-cell spatial phenotypes in high-content microscopy images. It segments cells, nuclei, pathogens and organelles with Cellpose, measures morphology, intensity, texture and colocalization features into a SQLite project, trains image and measurement classifiers, maps sequencing reads to wells and gRNAs, and estimates per-guide and per-gene effects by regression.
Autoencoder-based foundation model for 3D PDEs pretrained with online learning: a GPU-based synthetic-data solver generates training samples on the fly (no storage/IO bottleneck), scaling to an equivalent of hundreds of terabytes; learns transferable representations across heterogeneous physical systems with varying state variables and resolutions, and supports autoencoding, dynamics prediction via parameter-efficient LoRA fine-tuning, and generative modeling, with pretrained weights on HuggingFace (TU Munich Thuerey Group, 29+ stars, Apache 2.0, 2026)
Open-source package for atomistic simulations with machine learning and quantum chemical methods, home of the AIQM1/AIQM2 AI-enhanced quantum methods reaching beyond-DFT accuracy at semi-empirical cost, kernel ridge regression (KREG) potentials, native ANI neural-network implementations, Δ-learning workflows, and MDtrajNet trajectory models; usable as a Python library, via input files, or from the command line (Dral group / Xiamen University, 168+ stars, Apache 2.0, actively maintained)
Open-world multi-agent environment for autonomous scientific discovery where agents choose research directions, run experiments, discuss results, and publish papers into a shared literature (217+ stars, Apache 2.0)
The Cleavage Atlas collects proteolytic cleavage sites observed in public mass-spectrometry proteomics data (PRIDE/ProteomeXchange). Each identifier names a single peptide bond in a UniProtKB protein, independent of which protease cleaves it: the UniProtKB accession, an optional isoform number, and the P1 residue position (Schechter-Berger numbering on the full precursor sequence, Met1 = 1). The bond named is the one between residue P1 and residue P1+1.
The Digital Identity Optimization (DIO) and Ontology of Digital Identity (ODI) Reference Vocabulary is a domain-independent SKOS concept scheme that provides stable identifiers for concepts used to describe and analyze the representation, interpretation, trust, relationships, and reconstruction of digital identities across distributed information environments. It provides a shared, machine-readable terminology for researchers and practitioners working on digital identity, knowledge organization, semantic technologies, and AI-related systems.
'http://www.w3.org/ns/formats/Format' is a unique, top-level Class URI for specific file formats. Media types do not uniquely identify file formats (eg, serialization formats for various abstract concepts), because not all specific cases do have distinct media types, instances in this class can be used instead for proper identification. Although the URI-s themselves refer to non-informational resources (ie, they are abstract identifiers), they can be dereferenced. Dereferencing leads to informational resources in XHTML, RDF/XML, or Turtle equivalents, depending on content negotiation.
Schema elements for the [ioChem-BD](https://www.iochem-bd.org)
A RDF description of math
Represents molecular, electrochemical, computational, degradation, and provenance information relevant to organic redox flow battery research.
OntoIochem is an ontology designed to provide a semantic layer for the ioChem-BD platform and to support the consistent representation, discovery, and interoperability of computational chemistry data. Its development builds upon existing semantic resources, including the Gainesville ontology, which provided an important foundation for the representation of computational chemistry concepts and terminology. OntoIochem describes four main conceptual areas: the Molecular Concept, representing molecular structures and the properties that characterize them; the Calculation Concept, representing computational methodologies, software tools, calculation metadata, and derived results; the Publication Concept, representing information associated with scientific publications and research outputs, including data collections and manuscripts; and auxiliary concepts, such as units of measurement and values, which support the precise and machine-readable representation of scientific information. OntoIochem is designed to align with established computational chemistry ontologies and commonly used scientific terminologies, promoting semantic interoperability, reuse, and integration with existing resources and scientific data ecosystems. (from ontology)
The Structural Antibody Database (SAbDab2) aims to collect and annotate all publicly available, experimentally determined antibody structures. It is updated weekly, as new structures are added to the PDB, and includes single- and paired-chain antibodies, as well as their complexes. (from about page)
A RDF vocabulary of POSIX operations
ECMWF's unified open-source framework for building data-driven weather models, with modular packages for creating ML-ready training datasets from GRIB/NetCDF archives (anemoi-datasets), training graph neural networks and transformers (anemoi-training, powering ECMWF's AIFS forecasts), and running distributed inference (anemoi-inference), using config-driven recipes with proven HPC scalability (174+ stars, Apache 2.0, actively maintained, 2026)
Open-source site summarizing arXiv, bioRxiv and medRxiv papers, with paper Q&A, daily field alerts and reference checking (317+ stars, MIT License, 2023)
Scalable active-learning workflow engine for generating machine-learned interatomic potentials: orchestrates the full train–sample–label loop (QM reference calculations via CP2K/ORCA/GPAW up to coupled-cluster level, potential training, and enhanced-sampling MD via i-PI and PLUMED metadynamics) in Parsl-managed pipelines across local, HPC, and cloud nodes without databases or web servers; preconfigured Apptainer containers are production-ready for major supercomputers like LUMI and MeluXina (UAntwerp molmod group, 144+ stars, MIT License, actively maintained)
Multi-material, massively parallel Python physics engine for robot learning, built on Newton, MuJoCo Warp and Warp; simulates rigid robots, cloth, soft bodies, sand and liquids in one scene across thousands of parallel GPU worlds, with an MCP server so AI coding agents can build and run simulations (arXiv 2611.16312, 57+ stars, Apache 2.0, 2026)
Open-source framework for end-to-end retrosynthetic planning that combines multi-step search algorithms (MCTS, retro*, and retro-fallback) with wrappers for state-of-the-art single-step reaction models behind a standardized API, enabling benchmarking and rapid prototyping of synthesis-planning pipelines; used by RetroGFN, TANGO, and SimpRetro (arXiv 2310.19796, 217+ stars, MIT License, actively maintained)
British Antarctic Survey & Alan Turing Institute's probabilistic deep-learning system for seasonal Arctic sea-ice forecasting, training U-Net ensembles on CMIP6 historical simulations plus ERA5/OSI-SAF observations to predict monthly sea-ice concentration up to months ahead and outperforming SEAS5 at short lead times; the icenet-ai library is the pip-installable operational forecasting package with data download, training, and prediction pipelines (MIT License, 34+ stars, actively maintained)