WolframInstitute/BioDataLink
Find data sources, download verified files and read supported local measurements
Contributed By: Wolfram Institute
Find biological datasets, download a chosen file and read local measurements in Wolfram Language. Results retain identifiers, metadata and provenance.
Search
- Identifiers.org directory
- 774 namespaces, offline
- Pertpy dataset catalog
- 55 loader records, offline
- BioStudies studies and GEO Series
- online
Real data in the tutorials
- Single-cell expression
- 10x Genomics PBMC3k
- Protein structures
- PDB 4AKE and 1AKE
- Bacterial genome
- Borrelia burgdorferi B31
- Coronavirus genomes
- SARS-CoV-2, RaTG13 and SARS-CoV
- Circadian expression
- mouse liver, GEO GSE11923
Installation Instructions
To install this paclet in your Wolfram Language environment,
evaluate this code:
PacletInstall[ResourceObject["https://wolfr.am/1IzdEvymT"]]
To load the code after installation, evaluate this code:
Needs["WolframInstitute`BioDataLink`"]
Details
Search the bundled Identifiers.org directory and 55 Pertpy loader records offline. Listings are source metadata, not verified data imports.
Search public BioStudies studies or GEO Series online, one page at a time.
Download an explicit HTTPS file with a size limit and optional expected SHA-256. Requires curl 8.4 or later.
Read BioDataLink JSON as a cells-by-genes sparse matrix. Unlisted entries are zero; the format has no missing-measurement mask.
Read selected AnnData H5AD measurements with an explicitly configured Python interpreter, h5py and NumPy. This preserves selected axis metadata, not the complete AnnData object. Normalization is unspecified; no whole-file hash is computed.
Readers check structure without normalizing values or validating biological claims. Optional dependencies are installed separately.
Applications cover single-cell expression, protein structures, genomes and circadian rhythms. Real examples retain their sources, selections, licenses and checksums; synthetic examples are labeled.
The code is MIT licensed. Data keep their own terms: CC BY 4.0 for the registry, Iris and PBMC examples; CC0 for PDB structures; NCBI policies for sequences and GEO data.
After installation, load the package with Needs["WolframInstitute"].