the record
Every claim, sourced
If it is stated anywhere on this site or in anything posted about $bioLLM, it is listed here with the source it came from. Verified 2026-09-01.
| claim | source |
|---|---|
| The BioLLM framework for benchmarking single-cell foundation models | Qiu et al., Patterns 2025 — PMID 40843339 ↗ |
| scGPT — a generative foundation model for single-cell biology | Europe PMC ↗ |
| Geneformer — transfer learning on a human single-cell corpus | Europe PMC ↗ |
| Molecular formulae, weights and SMILES for indexed compounds | PubChem PUG-REST ↗ |
| Citations and abstracts for every indexed peptide | Europe PMC REST ↗ |
| Parent proteins for fragment peptides | UniProt REST ↗ |
| Experimental structures rendered from deposited coordinates | RCSB PDB ↗ |
| Registered clinical trials listed per peptide | ClinicalTrials.gov API v2 ↗ |
| Robinhood Chain is chain 4663 | Blockscout explorer ↗ |
| BPC-157 pentadecapeptide record | PubChem CID 9941957 ↗ |
| Thymosin β4, the parent of the TB-500 fragment | UniProt P62328 ↗ |
| Glucagon experimental structure | PDB 1GCN ↗ |
The database, in numbers
| peptides indexed | 100 |
|---|---|
| with an established sequence | 62 |
| independently mass-verified | 20 |
| citations fetched (Europe PMC) | 761 |
| PubChem compound records | 79 |
| entries with registered trials | 57 |
| experimental PDB structures | 6 |
| parent proteins (UniProt) | 46 |
| candidates published | 11 |