Real coverage stats across 7 modalities, refreshed continuously. Every record cites its source.
122,979,405 negative findings
Refresh cadence varies per source — shown on each card below.
39.1M of the small-molecule findings carry both compound structure and UniProt target — the substrate used in the validation paper to train 398 calibrated per-target bioactivity models. CRISPR negatives reproduce 98% across DepMap and BioGRID-ORCS (Analytics suite paper, Section 5).
Coverage depth differs by modality based on what exists in public databases. The tier label and counts below reflect the live state of the data.
85.2M records
Inactive compounds, failed selectivity panels, ADMET liabilities, and resistance mutations across all druggable target families.
e.g. “Inactive kinase inhibitors against EGFR”
37.6M records
Failed essentiality screens, non-performing guide RNAs, and ancestry-specific failures across human cell lines.
e.g. “Failed TP53 knockouts in primary T cells”
104K records
Terminated, withdrawn, and suspended clinical trials across all therapeutic areas, with intervention metadata and termination reasons.
e.g. “Terminated Phase 3 oncology trials in 2024”
3.6K records
Developability failures including aggregation, polyreactivity, expression, and immunogenicity. Plus terminated antibody trials.
Coverage note: Public-database coverage at the molecular level is thin by nature — most antibody developability data is proprietary. Enterprise tier supplementary materials extraction in progress.
e.g. “Antibodies with polyreactivity failures against PD-1”
18.8K records
Stability failures, oral bioavailability dead ends, immunogenicity issues, and proteolytic degradation for therapeutic peptides.
e.g. “Peptides with failed oral bioavailability”
824 records
Ternary complex failures, failed degradation, hook effect, and cell permeability issues for targeted protein degraders.
Coverage note: PROTAC discipline is ~7 years old; public coverage reflects the field's youth. Enterprise tier supplementary materials extraction in progress.
e.g. “Failed PROTACs targeting BTK”
144 records
Target engagement failures, delivery issues, hepatotoxicity, and immunogenicity for antisense oligonucleotides, siRNAs, and related modalities.
Coverage note: Most ASO/siRNA failure data is proprietary or in patents. Public coverage is thin. Enterprise tier supplementary materials extraction in progress.
e.g. “Failed ASOs against MALAT1”
Immunogenicity failures, animal efficacy failures, durability issues, and safety signals across vaccine platforms.
Status: Ingestion in progress. Expected live in v1.1.
Every record is queryable across modalities. Search for a target like EGFR and Nullary returns failed small molecules, failed CRISPR screens, failed antibody candidates, and terminated trials — all in one unified response with full citations.
Every record carries license metadata from its source so you can attribute or filter appropriately.
| Source | License | Attribution required | Commercial use |
|---|---|---|---|
| ChEMBL | CC-BY-SA 3.0 | Yes | Yes (share-alike) |
| PubChem BioAssay | Public Domain | No | Yes |
| BindingDB | CC-BY 3.0 | Yes | Yes |
| DepMap | CC-BY 4.0 | Yes | Yes |
| BioGRID ORCS | CC-BY 4.0 | Yes | Yes |
| GenomeCRISPR | CC-BY 4.0 | Yes | Yes |
| ClinicalTrials.gov (AACT) | Public Domain | No | Yes |
| EudraCT | Public Domain | No | Yes |
| Drugs@FDA | Public Domain | No | Yes |
| Thera-SAbDab | CC-BY 4.0 | Yes | Yes |
| SAbDab | CC-BY 4.0 | Yes | Yes |
| FLAb | Per-study, mostly CC-BY | Varies | Varies |
| OAS | CC-BY 4.0 | Yes | Yes |
| THPdb | Academic-use | Yes | Check terms |
| PepLife | Academic-use | Yes | Check terms |
| PROTAC-DB | CC-BY 4.0 | Yes | Yes |
| PROTACpedia | Academic-use | Yes | Check terms |
| AOBase | Academic-use | Yes | Check terms |
| Retraction Watch | CC-BY 4.0 | Yes | Yes |
Coverage depth varies by modality based on what exists in public databases. Small molecules and CRISPR have decades of accumulated public data. Newer modalities (PROTACs, oligonucleotides, ADCs) have thin public coverage because the fields themselves are younger and much of the failure data remains proprietary. The Enterprise tier extracts from supplementary materials, patents, and conference abstracts to expand coverage for under-served modalities.
Coverage as of 2026-05-25. The papers reference the 24–25 May 2026 snapshot; current numbers may differ slightly as ingestion continues.