整合靶点证据、药物管线、专利与交易开展BD评估
Coding
Target and Drug BD Assessment
Try itIntegrate target, pipeline, clinical, patent, scientific, and deal evidence for R&D and BD decisions.
What it does
Integrate target biology, disease rationale, drug pipeline, clinical, patent, scientific, regulatory, company, and transaction evidence for target or targeted-asset R&D and business-development decisions. Use for target initiation reviews, asset or modality opportunity screens, partnering theses, target landscape updates, and evidence-based go/no-go conditions.
The skill document
Assess Target and Drug BD Opportunities
Purpose
Turn a target or targeted-asset question into a traceable decision package for R&D initiation, portfolio review, partnering, licensing, or preliminary due diligence. Separate observed data, analyst inference, assumptions, and recommendations. Preserve sources, dates, geography, entity identity, coverage, and uncertainty.
This skill supports strategy. It is not medical advice, a valuation/fairness opinion, formal due diligence, or an FTO/legal opinion.
Inputs and decision frame
Confirm:
- target name, gene/protein/complex/isoform, aliases, species, and modality-specific target form;
- asset name/development code and sponsor if an asset is assessed;
- disease/indication, biomarker and patient segment;
- modality and mechanism of action;
- geography and language scope;
- analysis cutoff and decision horizon;
- decision: target initiation, modality selection, partner screen, asset diligence, competitor update, or go/no-go;
- audience and required depth/format;
- internal capabilities, constraints, comparator assets, and risk tolerance;
- confidential materials and permitted external services.
Normalize the target/entity first and show ambiguities before collecting downstream data. Do not merge related targets, complexes, ligands, paralogs or aliases without evidence.
MCP and evidence routing
Use supplied authoritative records when complete. For current retrieval, use only available/authorized global PatSnap MCPs:
| Evidence | MCP | Marketplace |
|---|---|---|
| Target/disease/epidemiology | target_disease | https://open.patsnap.com/marketplace/mcp-servers/target-disease |
| Drug identity/pipeline/milestones | drug_asset | https://open.patsnap.com/marketplace/mcp-servers/drug-asset |
| Trials and reported trial results | clinical_trials | https://open.patsnap.com/marketplace/mcp-servers/clinical-trials |
| Patent search | advanced_patent_search | https://open.patsnap.com/marketplace/mcp-servers/patent-search |
| Patent claims/family/status | patent_briefing | https://open.patsnap.com/marketplace/mcp-servers/patent-briefing |
| Translational evidence | scientific_translational_evidence | https://open.patsnap.com/marketplace/mcp-servers/scientific-translational-evidence |
| Guidelines/FDA-label evidence | regulatory_guidelines | https://open.patsnap.com/marketplace/mcp-servers/regulatory-guidelines |
| Pharmaceutical news | current_awareness | https://open.patsnap.com/marketplace/mcp-servers/current-awareness |
Use company/deal intelligence only when the current marketplace service and its callable tools are verified in the environment. Otherwise use company filings, press releases, regulatory filings, trial registries, or supplied exports and document the gap. Never invent legacy tool names, entity UUIDs, routes, or economic terms.
Evidence hierarchy
Prefer:
- official registries, regulator documents/labels, patents/registers, peer-reviewed primary research, trial publications, company filings and executed agreement text where available;
- structured PatSnap records with exact source/identifier/date;
- reputable secondary synthesis;
- news/web summaries as discovery leads or clearly labeled context.
For every material fact retain source ID, stable URL/identifier, event date, publication/retrieval date, entity, territory, field, value/unit, and verification status. Not found, not reported, not available, retrieval failed, and zero are different states.
Workflow
1. Target identity and biology
- normalized gene/protein/complex/isoform and identifiers;
- normal/tumor/disease expression by tissue/cell type with assay/dataset caveats;
- extracellular/intracellular location and accessibility;
- domain/structure and binding sites;
- pathway role and directionality;
- disease causal evidence: human genetics, perturbation, translational and pharmacological evidence;
- target dependence and responder/non-responder biology;
- safety liabilities, normal-tissue expression, redundancy and essentiality;
- resistance/escape and target heterogeneity;
- tractability by modality.
Do not convert expression correlation into causal validation. Score evidence classes separately.
2. Disease rationale and current care
- disease/subtype/stage and biomarker segment;
- incidence/prevalence/mortality with geography/year/source;
- current standard of care and guideline/regulatory date;
- outcomes, treatment gaps and safety/tolerability burden;
- diagnostic/biomarker pathway and testing feasibility;
- line of therapy and development positioning;
- target–disease mechanistic and clinical link;
- unmet need under a defined comparator.
3. Pipeline landscape
Retrieve all relevant assets within defined filters. Page until reported total is collected or disclose why not. Search by target/alias, drug/asset, modality, company, disease and geography as needed.
Normalize:
- drug/asset names and development codes;
- parent/active moiety and combinations;
- sponsor, originator, licensee and ownership change with effective date;
- modality/MoA, target form, binding site, payload/linker or construct;
- global and indication-specific highest phase;
- active/discontinued/terminated/unknown status;
- milestone source/date.
Deduplicate at both asset and indication-trial levels. A global phase does not prove activity in the target indication.
4. Clinical evidence
Keep registration facts and reported outcomes separate.
For material trials capture:
- registry ID and exact registry/source;
- phase, design, arms, comparator, blinding/randomization;
- population, biomarker, line, geography and enrollment;
- dose/route/schedule and combination;
- primary/secondary endpoints;
- recruitment/status and last verified date;
- reported result source, cutoff and analysis set;
- efficacy values with denominator/CI/time point;
- safety, discontinuation and dose-limiting toxicity;
- interpretation and limitations.
Do not fabricate or infer an NCT/registry ID. A trial-search result may require detail fetch/registry confirmation. Distinguish no result reported from negative result.
5. Competitor profiles
Select profiles by transparent criteria such as clinical maturity, mechanistic relevance, data strength, differentiation, strategic threat, deal activity or benchmark role—not a fixed Top 5.
For each asset:
- identity/sponsor/rights;
- modality, construct, binding site/MoA and differentiation;
- indications and phase/status by date;
- key trials/results;
- safety and development constraints;
- biomarker and patient-selection strategy;
- IP/family questions;
- transactions/partners;
- evidence strength, gaps and implications.
6. Patent landscape and IP questions
Search as applicable:
- target/antigen/epitope/MoA;
- modality, sequences, structures, payload/linker, conjugation and platform;
- composition, use, combination, biomarker, formulation, manufacturing and dosing;
- applicants/assignees/inventors, families and continuations;
- target jurisdictions and current status.
State query, dates, languages, jurisdictions, family/counting rule, retrieved/analyzed counts and sampling. A top-50 table is a sample, never full FTO coverage. Separate disclosure from claim coverage and verify material status in official registers. Route claim-level FTO to a dedicated jurisdiction-specific workflow and qualified counsel.
7. Transactions and ecosystem
Capture:
- announcement/effective/closing dates;
- licensor, licensee, acquirer, seller, partner and entity normalization;
- asset/target/modality/indication;
- transaction type;
- territory, development/commercial rights, options and sublicensing where disclosed;
- upfront, milestones, equity, royalties and total-potential value—each separately;
- disclosed versus undisclosed terms;
- original filing/announcement and secondary source;
- strategic interpretation and uncertainty.
Do not treat headline “up to” value as transaction value or revenue. Do not infer patent ownership/license scope from a press release alone.
8. Scientific and translational evidence
- foundational and recent primary research;
- human genetics, expression, mechanism and models;
- translational biomarkers and pharmacodynamic evidence;
- resistance/escape and combination rationale;
- conflicting/null/negative evidence;
- model-to-human translatability;
- publication type, date and limitations.
Report literature query and coverage. Total paper count is a database/search statistic, not research quality.
9. Regulatory, CMC and development feasibility
As relevant:
- approved precedents and regulator decisions by country/date;
- accelerated/conditional/orphan/fast-track pathways without assuming eligibility;
- endpoint and comparator precedent;
- diagnostic co-development;
- modality-specific manufacturing, analytics, stability, potency, delivery and scale risks;
- supply chain and control strategy;
- clinical operational feasibility and patient recruitment;
- evidence needed to de-risk.
10. Commercial, economic and financial framework
Only when requested and supported:
- addressable population funnel;
- treatment duration/penetration/scenario;
- benchmark price with country/year/gross-net caveat;
- market/payer/access constraints;
- cost/timeline ranges with source and uncertainty;
- scenario-based NPV/ROI assumptions, discount rate, probability, cash flows and sensitivity.
Do not present a single-point valuation or market forecast as fact. This skill does not provide a valuation opinion.
11. Decision synthesis
Assess separately:
- biological confidence;
- disease/biomarker rationale;
- modality–target fit;
- differentiated product profile;
- clinical feasibility and precedent;
- safety and resistance risk;
- CMC/manufacturing feasibility;
- patent/FTO uncertainty;
- competitive timing/crowding;
- partnering/rights availability;
- commercial/access attractiveness;
- organizational fit and next experiments.
State fact, inference, assumption and recommendation. Use confidence and sensitivity rather than opaque five-star totals.
Coverage rules
- Page every endpoint when reported total exceeds returned records; record pages, limit, retrieved count and deduped count.
- If pagination cannot complete, report the exact gap and do not claim “all.”
- Reconcile assets, indications, trials, patent publications/families and transactions under explicit units.
- Report per module: query/filter, source/database, total reported, retrieved, analyzed, excluded/deduplicated, cutoff and blind spots.
- Distinguish a genuine zero from unavailable/retrieval failure.
- Verify zero approved products with target, region and as-of date; do not generalize across indications/geographies.
Required output
- executive recommendation, decision type and confidence;
- target identity/biology and disease rationale;
- modality and pipeline landscape;
- evidence-based competitor profiles;
- clinical and translational evidence;
- preliminary patent landscape and IP questions;
- transactions/ecosystem and rights questions;
- regulatory/CMC/commercial considerations as applicable;
- differentiation options, risks and measurable go/no-go conditions;
- coverage summary, evidence gaps, sources and monitoring plan.
Use references/legacy-report-spec.md when the user requests a deep 21-section HTML presentation. The reference preserves the complete source information architecture but makes sections conditional and global.
Go/no-go conditions
Every condition requires:
- decision dimension;
- measurable threshold/event;
- evidence source/experiment;
- owner;
- decision date;
- GO, conditional GO, pause or NO-GO consequence;
- uncertainty and alternative interpretation.
Examples:
- target engagement demonstrated in a disease-relevant model by a prespecified assay;
- normal-tissue safety window meets threshold;
- biomarker enriches response reproducibly;
- competitor benchmark is exceeded on defined product attributes;
- manufacturing yield/potency/stability meets a target profile;
- target-country patent review identifies an actionable path;
- rights/partner availability meets stated commercial terms.
Do not use arbitrary P1/P2/P3 or star ratings without defined criteria.
Quality gates
- Target/entity identity and ambiguity are resolved or visible.
- Geography, indication, modality, cutoff and decision are explicit.
- Pipeline/trial/deal retrieval is fully paginated or caveated.
- Asset, indication, trial, patent and deal counts are deduplicated under stated units.
- Approval/trial/result claims have geography/status/date/source.
- Competitor selection/ranking criteria are explicit.
- Patent coverage and sampling are accurately described.
- Transaction parties, roles, rights, dates and disclosed/undisclosed economics are separate.
- Epidemiology/market/pricing/ROI assumptions include units, geography, date and sensitivity.
- Facts, inference, assumptions and recommendations are distinct.
- Each go/no-go condition is measurable and evidence-linked.
- No UUID, trial ID, transaction value, endpoint, patent, approval or source is fabricated.
Boundary
Recommend clinical, regulatory, CMC, commercial, financial, IP and legal specialist review for material decisions. Refresh dynamic pipeline, trial, status, transaction and guideline data at the decision date.
Related skills
药企药物研发辅助药物靶点筛选。参考 Open Targets Database skill 的 target-disease association 与 target prioritization 部分,构建药企研发场景下的候选靶点筛选能力。
Multi-region pharma intelligence that pulls regulatory, trial, patent, and biomedical data into source-cited answers.
Produce current, evidence-graded pharmaceutical and biotechnology analysis at the company, platform, pipeline, asset, indication, trial, competitive-landscape, catalyst, risk, and valuation levels. Use for biopharma diligence, investment-research-style reports, pipeline reviews, target or mechanism
Generate and prioritize experimentally testable target hypotheses from one or more small-molecule structures. Use when a user supplies SMILES or a compound library and asks which targets the molecules may modulate, how structural neighbors and SAR support the hypotheses, which biological and competitive evidence should be checked, or which compounds and targets should advance to orthogonal validation.
通过多维证据矩阵评估靶点药物并支持立项判断