Mosaic Pre-Clinical Intelligence

Get clinical trial pipeline for compounds targeting a gene. Returns compounds in clinical development with indications, trial phases, and status from ClinicalTrials.gov data.

Coverage-honest grid for ANY human gene symbol - the live demo of the mechanism. In this self-hosted package the grid resolves identity (HGNC) and structure (AlphaFold DB) live; the remaining axes are marked `queued` and are computed by the hosted service's dossier pipeline (getmosaic.dev). Every count carries a state; nothing unfetched is a zero.

Get compounds active against a specific drug target. Returns compounds with activity data (IC50, Ki, etc.), sorted by potency or, with sort='max_phase', approved and clinical drugs first. Unnamed research compounds read compound_name 'unnamed' with the ChEMBL id in chembl_id.

Get scientific papers about a specific drug target. Where the database carries PubTator3 gene annotations, only papers tagged with the target's NCBI Gene ID are served (`membership_basis` says which test applied), dated by the earliest of e-pub / issue / PubMed entry, papers naming the target in the title first.

Get a comprehensive intelligence dossier for a drug target. Returns UniProt biology, target scores, SAR summary, disease associations, validation evidence, pathways, PPIs, competitive landscape, clinical pipeline, and publication momentum. This is the primary tool for any target question.

Get the AlphaFold structural snapshot for a drug target. Returns AlphaFold model URLs (PDB / CIF / PAE), per-residue confidence summary (mean pLDDT, fractions of residues at high / confident / low confidence, disordered fraction), and protein length. Useful for SBDD scoping, disorder/IDR risk, and confidence-aware target triage. Also returns the fpocket pockets with both rankings (fpocket score and Druggability Score), which one selected the top pocket, and the pLDDT of the residues lining each pocket.

Get pathway context for a drug target. Shows which biological pathways the target participates in, other targets in the same pathways, and protein-protein interactions.

Candidate resistance-bypass / escape targets for a given target. From a deterministic keyword pass over the literature (resistance_relations — GLiREL has no resistance edge type), surfaces targets co-mentioned with the query target in resistance-context abstracts, ranked by a drugability-gap score (strong resistance evidence, low development activity). Hypothesis generator, not evidence — every row carries its source snippet.

Find synthetic-lethal *whitespace*: targets functionally coupled to a developed (drugged) target but themselves undeveloped. For an anchor target with chemical matter, surfaces partners coupled to it by DepMap co-essentiality, STRING protein-protein interactions and/or Reactome pathways. Returns TWO lists, deliberately not merged: - `candidates` — partners Mosaic covers, filtered to < 5 patents and no compound past phase 0. Those counts are measured, so a whitespace claim is supportable. Ranked by `whitespace_score`. **Read `clinical_absence_caveat`**: it is set when a partner carries substantial chemistry yet nothing clinical, which usually means our compound set is incomplete rather than that the world has no clinical interest. - `coupled_unassessed` — partners outside the curated universe. They are coupled to the anchor, but their competitive status is UNKNOWN rather than zero, so they carry null counts and no whitespace_score, and are ranked by `coupling_strength` alone. Leads, not evidence. Do not describe them as uncontested. Scope: DepMap co-essentiality IS ingested, but most rows do not use it. `co_functionality_basis` is one of `depmap_coessentiality` (measured co-essentiality), `ppi_pathway_proxy` (protein interaction AND a shared pathway), `ppi_only` (protein interaction alone, pathways checked and none shared — the weakest, and on a KRAS anchor the most common), or `ppi_pathways_unassessed` (interaction only, and pathways COULD NOT be checked because the partner is outside the curated universe — those rows also carry `shared_pathways: null`, not 0). `basis_note` gives the split. A proxy row is a hypothesis about functional coupling, not measured synthetic lethality. `sufficient_for_absence_claim` is false unless most of the coupled set was assessable; `assessed_fraction` reports how much was. `lineage` is still a label only — it does not filter. Hypothesis generator.

Get the 1-hop neighborhood around a drug target. Returns the entities DIRECTLY connected to the target — compounds, diseases, pathways, patent-holding organizations, interacting proteins — as nodes and edges. This is a star centered on the target (independent 1-hop lookups), not a multi-hop graph traversal: it shows the target's immediate context, not paths that hop through intermediate entities.

Assay precedent for a drug target: what has been tried, in what model system. Returns literature-derived experimental precedent — genetic (CRISPR/siRNA/KO), in vivo (mouse/rat/xenograft), pharmacological (inhibitor/SAR), and clinical- trial evidence — re-classified per paper at query time. A paper may carry several methods; counts are lower bounds, not exhaustive. Papers that name the target in their title are flagged high-confidence and lead the ranked exemplars. Outcomes are NOT auto-graded: open the linked exemplar papers to read what actually happened. Also returns DepMap essentiality + AlphaMissense pathogenicity when available. A target with no precedent returns an explicit no_precedent_in_corpus status (corpus coverage, not confirmed absence).