Sequence first, epitope second. Start from a known target, recover antibody/antigen sequences and SEQ IDs, then evaluate patent/public evidence and optional epitope disclosure.

This module is the sequence-first IP scouting workspace for known-target antibody programs. It turns scattered patent, sequence, literature, epitope, and structure evidence into a single IP-aware research brief. The core object is sequence: therapeutic antibody sequence, antigen sequence, CDR/SEQ ID, clone, patent-derived sequence evidence, and accession-level sequence leads. Epitope is a downstream interpretation layer after sequence recovery. It is not an FTO opinion or legal opinion; it is the research operating layer before claim-chart, design-around, FTO packet preparation, or downstream binder work.

Target

Canonical target name, such as FCRL5, DLL3, BCMA, or a gene/protein symbol.

Aliases

Synonyms, older names, CD markers, UniProt-style names, and project shorthand. These expand the query pack.

Therapeutic / exclusion context

The business boundary: therapeutic antibody or CAR binder focus, and exclusions such as IHC, ELISA, flow cytometry, diagnostic kits, or research reagents.

API credentials

Optional LENS_API_TOKEN enables Lens Patent API recall. Without it, the page still runs the no-key public connectors.

Manual evidence

Patent family IDs, claim excerpts, clone names, PDFs, sequence listings, or counsel notes can be pasted into Evidence notes after the automated pass.

In-platform search results

Grouped source results from live APIs: Lens Patent API when configured, Europe PMC, NCBI patent sequences, IEDB B-cell epitope records, and RCSB PDB.

Sequence-first evidence table

Target, recovered sequence or sequence lead, therapeutic antibody label, optional epitope disclosure, and sourcing basis. The table can be downloaded as CSV.

Evidence notes

Editable working notes for patent families/claims, sequence evidence, epitope evidence, and next-action decisions.

Brief

A copyable Markdown brief that combines the query pack, source hits, table rows, and human enrichment notes.

FTO handoff packet

A research packet for patent counsel: sequence leads, source basis, claim clues, family/legal-status fields, and explicit boundary notes. It is not a legal conclusion.

1. 1. Open the page from Workspace → Sequence Patent Scout

   Open the page from Workspace → Sequence Patent Scout. The module page is the operating surface; this walkthrough is the explanation layer.

2. 2. Enter the canonical target and aliases

   Enter the canonical target and aliases. Include CD markers, older names, spelling variants, FcRH-style names, and internal shorthand because patent and sequence records often use inconsistent naming.

3. 3. State the therapeutic boundary

   State the therapeutic boundary. Write down that the search is focused on therapeutic antibodies, CAR binders, or antigen-binding domains, and explicitly exclude diagnostic antibodies, IHC, ELISA, flow cytometry clones, kits, and reagent-only uses.

4. 4. Run the AI-first scout

   Run the AI-first scout. The platform performs the first pass automatically: query-pack construction, API recall, source grouping, sequence/SEQ ID recovery, sequence-centric evidence normalization, and table generation.

5. 5. Review source groups from strongest sequence signal to weakest

   Review source groups from strongest sequence signal to weakest. NCBI Patent Sequence can expose protein/nucleotide FASTA or accession-level sequence leads. Lens Patent API, when configured, adds patent title, applicant/owner, family/legal-status, and sequence-listing metadata. Europe PMC gives literature and public biomedical context. IEDB only answers whether epitope records exist. RCSB adds structure evidence.

6. 6. Read the sequence-first evidence table

   Read the sequence-first evidence table. Start with the Sequence / SEQ ID column, not the epitope column. A full FASTA, VH/VL/CDR record, antigen sequence, clone, or accession is valuable even when no epitope is disclosed.

7. 7. Classify each row

   Classify each row. Keep rows with actual therapeutic sequence evidence; keep accession-level leads for follow-up; mark diagnostic/reagent-only hits as exclusions; mark weak literature-only hits as context rather than sequence evidence.

8. 8. Use Add evidence on relevant rows

   Use Add evidence on relevant rows. Patent-like hits go to patent family/claims notes; patent sequence hits go to sequence evidence; IEDB hits go to epitope evidence. Manual work starts after the AI-first pass: enrich missing context, correct weak labels, and add source-specific notes.

9. 9. Decide the epitope path only after sequence recovery

   Decide the epitope path only after sequence recovery. If epitope or competition/blocking is explicitly disclosed, prepare claim-chart and design-around discussion. If epitope is absent, continue with the sequence by using Interface, structure prediction, and possible wet-lab competition assays to infer footprint.

10. 10. Download CSV or copy the brief

    Download CSV or copy the brief. Use the CSV for team review and the brief for IP counsel or downstream platform owners. Keep the output framed as an evidence packet, not an FTO opinion.

• POST /api/v1/patent-epitope-scout/search is the aggregation endpoint.
• Live no-key connectors: Europe PMC REST API, NCBI E-utilities for protein/nuccore patent sequences, IEDB IQ-API B-cell search, and RCSB Search/Data APIs.
• Lens Patent API auto-runs when LENS_API_TOKEN is configured, adding patent title, applicant/owner, legal status, family size, and sequence_listing metadata.
• Needs-key or bulk connectors still visible for later expansion: Lens PatSeq/PatSeq Finder, EPO OPS, USPTO ODP, WIPO sequence listings, SAbDab/Thera-SAbDab, and clinical/approved-product sources.
• The current table uses deterministic source-normalization heuristics to surface sequence/SEQ ID leads first, then infer antibody labels and epitope disclosure from source hits. This is AI-first triage; humans should enrich and correct before final claim interpretation.
• Sequence display is opportunistic: NCBI EFetch can return FASTA for many patent-sequence records; when full sequence is unavailable, the table preserves accession/SEQ ID as a follow-up sequence lead.
• FTO planning should be added as a separate downstream layer: product element definition, jurisdiction scope, claim mapping, family/legal-status review, expiry/term check, and counsel-reviewed risk classification.
• Boundary: public APIs have rate limits and incomplete coverage. Patents can hide critical sequence or claim context in PDFs, image files, file wrappers, or sequence listings not yet parsed by the live connector.

• A patent sequence hit is not automatically a therapeutic antibody. It may be a construct, antigen, linker, vector, or unrelated sequence within the same patent.
• Absence of an epitope hit is not evidence that no epitope is claimed. It also does not make the row useless: a recovered sequence or sequence lead is still a valuable downstream input.
• Diagnostic clones often share target names with therapeutic antibodies; keep the exclusion context explicit.
• Claims can protect sequence, CDR identity, epitope, competition, composition, or method-of-treatment. The table is a starting point, not a final claim chart.
• Do not collapse sequence scouting and FTO into one conclusion. The platform can prepare an evidence packet and risk labels; counsel must own the formal FTO opinion.