@bgicli/bgicli 2.1.1 → 2.2.0

package/data/workflows/clinicaltrials-landscape/SKILL.md

@@ -0,0 +1,257 @@
+---
+id: clinicaltrials-landscape
+name: ClinicalTrials.gov Disease Landscape Scanner
+category: literature
+short-description: "Query ClinicalTrials.gov API v2 to map the clinical trial landscape for any disease area by mechanism, phase, and sponsor."
+detailed-description: "Programmatically query the free ClinicalTrials.gov API v2 to pull all active clinical trials for a disease area, classify by therapeutic mechanism of action, and generate competitive landscape visualizations. Supports any disease with pre-built configs for IBD (Crohn's, UC). Generic mode classifies by intervention type when no disease config exists. Supports filtering by mechanism, phase, sponsor, and status. Exports structured CSVs, publication-quality plots, and pickle objects for downstream analysis. No API key required."
+starting-prompt: Show me the current clinical trial landscape for IBD
+---
+
+# ClinicalTrials.gov Disease Landscape Scanner
+
+## When to Use This Skill
+
+- **Map competitive landscape** across therapeutic mechanisms for any disease
+- **Track specific mechanism classes** (e.g., anti-IL23, anti-TL1A, JAK inhibitors)
+- **Identify sponsors** and their pipeline positions by phase
+- **Phase distribution analysis** for business development diligence
+- **Pipeline monitoring** for a specific sponsor's disease portfolio
+- **Pre-built disease configs** available (IBD with 14 mechanism classes); generic mode for any other disease
+
+**Do NOT use for:**
+- Detailed single-trial protocol analysis
+- Efficacy/safety comparisons (requires literature review skill)
+
+---
+
+## Installation
+
+| Software | Version | License | Commercial Use | Installation |
+|----------|---------|---------|----------------|-------------|
+| pandas | ≥1.3 | BSD-3 | ✅ Permitted | `pip install pandas` |
+| requests | ≥2.25 | Apache-2.0 | ✅ Permitted | `pip install requests` |
+| numpy | ≥1.20 | BSD-3 | ✅ Permitted | `pip install numpy` |
+| plotnine | ≥0.10 | MIT | ✅ Permitted | `pip install plotnine` |
+| plotnine-prism | ≥0.3 | MIT | ✅ Permitted | `pip install plotnine-prism` |
+| seaborn | ≥0.11 | BSD-3 | ✅ Permitted | `pip install seaborn` |
+| matplotlib | ≥3.4 | PSF | ✅ Permitted | `pip install matplotlib` |
+| reportlab | ≥3.6 | BSD | ✅ Permitted | `pip install reportlab` |
+| pyyaml | ≥5.0 | MIT | ✅ Permitted | `pip install pyyaml` |
+
+```bash
+pip install pandas requests numpy plotnine plotnine-prism seaborn matplotlib reportlab pyyaml
+```
+
+**System requirements:** Internet connection for ClinicalTrials.gov API calls.
+
+---
+
+## Inputs
+
+**Required:**
+- **Disease / condition terms** — list of conditions to search ClinicalTrials.gov
+
+**Optional:**
+- **Disease config** — pre-built config ID (e.g., `"ibd"`) for mechanism taxonomy, or `None` for generic
+- **Mechanism filter** — e.g., "Anti-IL-23 (p19)", "Anti-TL1A", "JAK Inhibitor"
+- **Sponsor filter** — e.g., "Takeda", "AbbVie"
+- **Status filter** — Default: all active (Recruiting + Active not recruiting + Not yet recruiting)
+- **Phase filter** — Phase 1, 2, 3, 4
+
+---
+
+## Outputs
+
+**Visualizations (PNG + SVG):**
+- `landscape_overview.png/.svg` — 6-panel landscape figure (300 DPI)
+  - Mechanism × Phase heatmap, top sponsors, phase stacked bars, mechanism counts, timeline, sponsor type
+- `landscape_supplementary.png/.svg` — 4-panel supplementary figure
+  - Top 15 countries, study design by phase, enrollment distribution, phase transition funnel
+
+**Results (CSV):**
+- `trials_all.csv` — All trials with 46 columns (mechanism, phase, sponsor, geography, study design, arms, endpoints, eligibility, regulatory)
+- `trials_by_mechanism.csv` — Mechanism × phase cross-tabulation
+- `trials_by_sponsor.csv` — Sponsor summary with trial counts
+- `trials_filtered.csv` — Filtered subset (if mechanism/sponsor filter applied)
+
+**Reports:**
+- `landscape_report.pdf` — Publication-quality PDF with 24 sections: executive summary, mechanism deep-dives, geographic landscape, study design, phase transition funnel, endpoint comparison, combination therapies, biosimilar assessment, whitespace analysis, and more
+- `landscape_report.md` — Markdown version with identical 24-section structure
+
+**Analysis objects (Pickle):**
+- `analysis_object.pkl` — Complete landscape for downstream use
+  - Load with: `import pickle; obj = pickle.load(open('analysis_object.pkl', 'rb'))`
+  - Contains: trials_df (46 columns), mechanism/phase/sponsor distributions, geographic stats, design stats, parameters
+
+---
+
+## Clarification Questions
+
+1. **Data Source** (ASK THIS FIRST):
+   - This skill queries the ClinicalTrials.gov API v2 directly (free, no key needed).
+   - **Use live API data?** (recommended, ~30 seconds)
+   - **Or use cached demo data?** Pre-loaded IBD landscape snapshot for quick demo
+
+2. **Disease Area:**
+   - Which disease area to analyze?
+     - a) IBD (Inflammatory Bowel Disease) — pre-built config with 14 mechanism classes
+     - b) Oncology (generic intervention-type classification)
+     - c) Autoimmune / Rheumatology (generic classification)
+     - d) Other (specify disease and condition terms)
+
+3. **Scope** *(if IBD selected)*:
+   - Which conditions?
+     - a) All IBD (Crohn's, UC, and IBD unspecified) — recommended
+     - b) Crohn's Disease only
+     - c) Ulcerative Colitis only
+   - *(If other disease)* — Provide list of condition search terms
+
+4. **Focus:**
+   - Any mechanism or sponsor to highlight?
+     - *(IBD)* a) Anti-IL-23 — recommended for demo | b) Anti-TL1A | c) All mechanisms
+     - *(Other)* Specify or skip highlighting
+
+---
+
+## Standard Workflow
+
+🚨 **MANDATORY: USE SCRIPTS EXACTLY AS SHOWN - DO NOT WRITE INLINE CODE** 🚨
+
+**Step 1 — Load config and query ClinicalTrials.gov:**
+```python
+import sys; sys.path.insert(0, ".")
+from scripts.disease_config import load_disease_config, get_default_conditions
+from scripts.query_clinicaltrials import query_trials
+
+# Load disease config (use "ibd" for IBD, or None for generic)
+config = load_disease_config("ibd")
+
+# Get conditions from config or specify manually
+conditions = get_default_conditions(config) or ["Crohn's Disease", "Ulcerative Colitis", "Inflammatory Bowel Disease"]
+
+raw_trials = query_trials(
+    conditions=conditions,
+    statuses=["RECRUITING", "ACTIVE_NOT_RECRUITING", "ENROLLING_BY_INVITATION", "NOT_YET_RECRUITING"],
+)
+```
+**✅ VERIFICATION:** `"✓ Retrieved {N} trials from ClinicalTrials.gov"`
+
+**Step 2 — Classify and compile:**
+```python
+from scripts.classify_mechanisms import classify_all
+from scripts.compile_trials import compile_trials
+
+classified = classify_all(raw_trials, config=config)
+trials_df = compile_trials(classified, output_dir="landscape_results")
+```
+**DO NOT write inline classification code. The script loads mechanism taxonomy from config.**
+
+**✅ VERIFICATION:** `"✓ Trial data compiled successfully!"`
+
+**Step 3 — Generate visualizations:**
+```python
+from scripts.generate_landscape_plots import generate_landscape_plots
+
+generate_landscape_plots(
+    trials_df,
+    output_dir="landscape_results",
+    highlight_mechanism="Anti-IL-23 (p19)",  # or None for no highlight
+    highlight_sponsor=None,                   # or "Takeda" to highlight
+    config=config,
+)
+```
+🚨 **DO NOT write inline plotting code. The script handles all 6 panels + PNG/SVG export.** 🚨
+
+**✅ VERIFICATION:** `"✓ All landscape visualizations generated successfully!"`
+
+**Step 4 — Export results:**
+```python
+from scripts.export_all import export_all
+
+export_all(
+    trials_df,
+    parameters={
+        "conditions": conditions,
+        "statuses": ["RECRUITING", "ACTIVE_NOT_RECRUITING", "ENROLLING_BY_INVITATION", "NOT_YET_RECRUITING"],
+        "highlight_mechanism": "Anti-IL-23 (p19)",
+    },
+    output_dir="landscape_results",
+    config=config,
+)
+```
+**DO NOT write custom export code. Use export_all().**
+
+**✅ VERIFICATION:** `"=== Export Complete ==="`
+
+---
+
+## ⚠️ CRITICAL — DO NOT:
+
+- ❌ **Write inline classification code** → **STOP: Use `classify_all()` from scripts**
+- ❌ **Write inline plotting code (ggplot, plt, sns)** → **STOP: Use `generate_landscape_plots()`**
+- ❌ **Write custom export code** → **STOP: Use `export_all()`**
+- ❌ **Try to scrape ClinicalTrials.gov HTML** → **Use the API via `query_trials()`**
+
+---
+
+## ⚠️ IF SCRIPTS FAIL — Script Failure Hierarchy:
+
+1. **Fix and Retry (90%)** — Install missing package, re-run script
+2. **Modify Script (5%)** — Edit the script file itself, document changes
+3. **Use as Reference (4%)** — Read script, adapt approach, cite source
+4. **Write from Scratch (1%)** — Only if genuinely impossible, explain why
+
+**NEVER skip directly to writing inline code without trying the script first.**
+
+---
+
+## Common Issues
+
+| Error | Cause | Solution |
+|-------|-------|----------|
+| **ConnectionError / Timeout** | ClinicalTrials.gov unreachable | Check internet connection; retry after 30 seconds |
+| **HTTP 429 Too Many Requests** | Rate limit exceeded | Increase `RATE_LIMIT_DELAY` in query_clinicaltrials.py |
+| **ModuleNotFoundError: plotnine** | Missing visualization package | `pip install plotnine plotnine-prism` |
+| **Empty results (0 trials)** | Overly restrictive filters | Broaden condition/status/phase filters |
+| **Many "Unclassified" mechanisms** | No disease config or new drugs | Use a disease config (e.g., `"ibd"`) or update `disease_configs/*.yaml` |
+| **SVG export failed** | Missing SVG backend | Normal — PNG is always generated as fallback |
+| **Sponsor name variants** | Same company, different names | Update `SPONSOR_NORMALIZATION` in compile_trials.py |
+| **ModuleNotFoundError: yaml** | Missing pyyaml | `pip install pyyaml` |
+
+---
+
+## Interpretation Guidelines
+
+- **Mechanism classification** is based on intervention names and descriptions — some trials with vague descriptions (e.g., "Study Drug") will be classified as "Other Biologic" or "Unclassified"
+- **Phase 2/3** indicates a combined Phase 2/3 study design
+- **Sponsor normalization** groups subsidiaries under parent company (e.g., Millennium → Takeda)
+- **Industry vs Academic** based on ClinicalTrials.gov `leadSponsor.class` field
+- The landscape reflects **registered trials**, not all pipeline programs (pre-IND programs won't appear)
+- **Disease configs** provide curated mechanism taxonomies; without config, classification uses generic intervention types
+
+---
+
+## Suggested Next Steps
+
+1. **Deep-dive a mechanism** — Use `literature-preclinical` to review mechanism biology
+2. **Track a sponsor's full pipeline** — Use `development-landscape` for broader pipeline view
+3. **Biomarker analysis** — Use `lasso-biomarker-panel` to identify response biomarkers from trial data
+4. **Export to presentation** — Use landscape_report.md and plots for stakeholder review
+
+---
+
+## Related Skills
+
+- `development-landscape` — Broader, multi-source pipeline landscape for any target
+- `literature-preclinical` — Literature review for mechanism biology
+- `lasso-biomarker-panel` — Biomarker discovery from expression data
+
+---
+
+## References
+
+- ClinicalTrials.gov API v2: https://clinicaltrials.gov/data-api/api
+- ClinicalTrials.gov: https://clinicaltrials.gov/
+- See `references/api-parameters.md` for full API parameter reference
+- See `references/mechanisms.md` for mechanism taxonomy details
+- See `references/output-schema.md` for output column definitions

package/data/workflows/clinicaltrials-landscape/references/api-parameters.md

@@ -0,0 +1,181 @@
+# ClinicalTrials.gov API v2 Reference
+
+## Endpoint
+
+```
+GET https://clinicaltrials.gov/api/v2/studies
+```
+
+Free, no authentication required. Rate limit ~50 requests/minute.
+
+## Query Parameters
+
+| Parameter              | Type    | Description                             | Example                                     |
+| ---------------------- | ------- | --------------------------------------- | ------------------------------------------- |
+| `query.cond`           | string  | Condition/disease search                | `"Crohn's Disease" OR "Ulcerative Colitis"` |
+| `query.intr`           | string  | Intervention search                     | `risankizumab`                              |
+| `query.spons`          | string  | Sponsor name search                     | `Takeda`                                    |
+| `query.locn`           | string  | Location search                         | `United States`                             |
+| `query.term`           | string  | General search                          | `anti-IL23 IBD`                             |
+| `filter.overallStatus` | string  | Comma-separated status filter           | `RECRUITING,ACTIVE_NOT_RECRUITING`          |
+| `filter.phase`         | string  | Comma-separated phase filter            | `PHASE2,PHASE3`                             |
+| `pageSize`             | integer | Results per page (max 1000)             | `1000`                                      |
+| `pageToken`            | string  | Pagination token from previous response |                                             |
+| `countTotal`           | boolean | Include total count in response         | `true`                                      |
+| `format`               | string  | Response format                         | `json` or `csv`                             |
+| `sort`                 | string  | Sort order                              | `LastUpdatePostDate:desc`                   |
+
+## Status Values
+
+| API Value                 | Meaning                          |
+| ------------------------- | -------------------------------- |
+| `RECRUITING`              | Actively recruiting participants |
+| `NOT_YET_RECRUITING`      | Not yet started recruiting       |
+| `ACTIVE_NOT_RECRUITING`   | Ongoing but not enrolling        |
+| `ENROLLING_BY_INVITATION` | Enrolling by invitation only     |
+| `COMPLETED`               | Study completed                  |
+| `TERMINATED`              | Stopped early                    |
+| `WITHDRAWN`               | Withdrawn before enrollment      |
+| `SUSPENDED`               | Temporarily paused               |
+
+## Phase Values
+
+| API Value      | Display        |
+| -------------- | -------------- |
+| `EARLY_PHASE1` | Phase 1        |
+| `PHASE1`       | Phase 1        |
+| `PHASE2`       | Phase 2        |
+| `PHASE3`       | Phase 3        |
+| `PHASE4`       | Phase 4        |
+| `NA`           | Not Applicable |
+
+Note: A study can have multiple phases (e.g., `["PHASE2", "PHASE3"]` for Phase
+2/3 studies).
+
+## Response Structure
+
+```json
+{
+  "totalCount": 1234,
+  "nextPageToken": "...",
+  "studies": [
+    {
+      "protocolSection": {
+        "identificationModule": {
+          "nctId": "NCT...",
+          "briefTitle": "...",
+          "officialTitle": "..."
+        },
+        "statusModule": {
+          "overallStatus": "RECRUITING",
+          "startDateStruct": { "date": "2023-01" },
+          "primaryCompletionDateStruct": { "date": "2025-06" },
+          "completionDateStruct": { "date": "2025-12" }
+        },
+        "sponsorCollaboratorsModule": {
+          "leadSponsor": { "name": "...", "class": "INDUSTRY" },
+          "collaborators": [{ "name": "...", "class": "OTHER" }]
+        },
+        "conditionsModule": { "conditions": ["Crohn's Disease"] },
+        "armsInterventionsModule": {
+          "interventions": [
+            { "type": "DRUG", "name": "...", "description": "..." }
+          ],
+          "arms": [
+            {
+              "label": "Treatment",
+              "type": "EXPERIMENTAL",
+              "description": "..."
+            }
+          ]
+        },
+        "designModule": {
+          "studyType": "INTERVENTIONAL",
+          "phases": ["PHASE3"],
+          "enrollmentInfo": { "count": 500 },
+          "designInfo": {
+            "allocation": "RANDOMIZED",
+            "interventionModel": "PARALLEL",
+            "maskingInfo": { "masking": "DOUBLE" }
+          }
+        },
+        "contactsLocationsModule": {
+          "locations": [
+            { "country": "United States" },
+            { "country": "Germany" }
+          ]
+        },
+        "outcomesModule": {
+          "primaryOutcomes": [
+            { "measure": "Clinical remission", "timeFrame": "52 weeks" }
+          ]
+        },
+        "eligibilityModule": {
+          "minimumAge": "18 Years",
+          "maximumAge": "80 Years",
+          "sex": "ALL"
+        },
+        "oversightModule": {
+          "isFdaRegulatedDrug": true,
+          "oversightHasDmc": true
+        },
+        "descriptionModule": {
+          "briefSummary": "A study of..."
+        }
+      }
+    }
+  ]
+}
+```
+
+## Extracted Modules
+
+The skill extracts data from the following API response modules:
+
+| Module                       | Fields Extracted                                                                           | Purpose                                  |
+| ---------------------------- | ------------------------------------------------------------------------------------------ | ---------------------------------------- |
+| `identificationModule`       | nctId, briefTitle, officialTitle                                                           | Trial identification                     |
+| `statusModule`               | overallStatus, startDateStruct, completionDateStruct                                       | Status and timeline                      |
+| `sponsorCollaboratorsModule` | leadSponsor (name, class), collaborators                                                   | Sponsor and partnerships                 |
+| `conditionsModule`           | conditions                                                                                 | Disease indications                      |
+| `armsInterventionsModule`    | interventions (type, name, description), arms (label, type)                                | Drug classification, comparator analysis |
+| `designModule`               | phases, enrollmentInfo, studyType, designInfo (allocation, interventionModel, maskingInfo) | Phase, enrollment, study design          |
+| `contactsLocationsModule`    | locations[].country                                                                        | Geographic analysis                      |
+| `outcomesModule`             | primaryOutcomes (measure, timeFrame)                                                       | Endpoint comparison                      |
+| `eligibilityModule`          | minimumAge, maximumAge, sex                                                                | Patient population analysis              |
+| `oversightModule`            | isFdaRegulatedDrug, oversightHasDmc                                                        | Regulatory signals                       |
+| `descriptionModule`          | briefSummary                                                                               | Secondary mechanism classification       |
+
+## Sponsor Classes
+
+| Value       | Meaning                        |
+| ----------- | ------------------------------ |
+| `INDUSTRY`  | Pharmaceutical/biotech company |
+| `NIH`       | National Institutes of Health  |
+| `FED`       | Other US Federal agency        |
+| `OTHER`     | Academic, hospital, or other   |
+| `OTHER_GOV` | Non-US government              |
+| `NETWORK`   | Clinical network               |
+| `INDIV`     | Individual investigator        |
+
+## Pagination
+
+Results are paginated. Use `nextPageToken` from the response in subsequent
+requests:
+
+```python
+# First request
+response = requests.get(BASE_URL, params={"query.cond": "IBD", "pageSize": 1000})
+data = response.json()
+
+# Next page
+next_token = data.get("nextPageToken")
+if next_token:
+    response = requests.get(BASE_URL, params={..., "pageToken": next_token})
+```
+
+## Rate Limiting
+
+- ~50 requests per minute per IP
+- Use 1-2 second delay between paginated requests
+- No authentication required

package/data/workflows/clinicaltrials-landscape/references/mechanisms.md

@@ -0,0 +1,141 @@
+# Therapeutic Mechanism Taxonomy
+
+Classification guide for clinical trial interventions by mechanism of action.
+This reference documents the IBD taxonomy. Other disease configs are in
+`disease_configs/`.
+
+## Mechanism Classes
+
+### Anti-TNF
+
+- **Biology:** Block tumor necrosis factor alpha (TNFα), a pro-inflammatory
+  cytokine
+- **Approved drugs:** Infliximab (Remicade), Adalimumab (Humira), Certolizumab
+  pegol (Cimzia), Golimumab (Simponi)
+- **Biosimilars:** CT-P13 (Remsima/Inflectra), SB2 (Renflexis), many others
+- **Notes:** First biologic class for IBD; mature market with extensive
+  biosimilar competition
+
+### Anti-IL-12/23 (p40)
+
+- **Biology:** Block shared p40 subunit of IL-12 and IL-23
+- **Approved drugs:** Ustekinumab (Stelara)
+- **Notes:** Targets both Th1 (IL-12) and Th17 (IL-23) pathways; broader
+  mechanism than p19-selective
+
+### Anti-IL-23 (p19)
+
+- **Biology:** Selectively block IL-23 via p19 subunit; spares IL-12/Th1 pathway
+- **Key drugs:** Risankizumab (Skyrizi, AbbVie), Guselkumab (Tremfya, J&J),
+  Mirikizumab (Omvoh, Eli Lilly), Brazikumab (AstraZeneca)
+- **Notes:** High-activity class; multiple Phase 3 programs. More selective than
+  ustekinumab
+
+### JAK Inhibitor
+
+- **Biology:** Small molecules blocking Janus kinase signaling (JAK1, JAK2,
+  JAK3, TYK2)
+- **Key drugs:** Tofacitinib (Xeljanz, Pfizer, pan-JAK), Upadacitinib (Rinvoq,
+  AbbVie, JAK1-selective), Filgotinib (Jyseleca, Galapagos/Gilead,
+  JAK1-selective), Deucravacitinib (BMS, TYK2-selective)
+- **Notes:** Oral small molecules; safety monitoring required (cardiovascular,
+  infections, malignancy)
+
+### Anti-Integrin
+
+- **Biology:** Block leukocyte trafficking to gut via integrin receptors
+- **Key drugs:** Vedolizumab (Entyvio, **Takeda**, gut-selective α4β7),
+  Natalizumab (Tysabri, α4-integrin), Etrolizumab (Roche, α4β7/αEβ7),
+  Ontamalimab (anti-MAdCAM-1)
+- **Takeda relevance:** Vedolizumab (Entyvio) is Takeda's flagship IBD product
+- **Notes:** Gut-selective mechanism attractive for safety profile
+
+### S1P Modulator
+
+- **Biology:** Sequester lymphocytes in lymph nodes by modulating
+  sphingosine-1-phosphate receptors
+- **Key drugs:** Ozanimod (Zeposia, BMS), Etrasimod (Velsipity, Pfizer/Arena)
+- **Notes:** Oral, non-JAK mechanism; growing class
+
+### Anti-TL1A
+
+- **Biology:** Block TNF-like ligand 1A (TL1A/TNFSF15), involved in mucosal
+  inflammation and fibrosis
+- **Key drugs:** Duvakitug (PRA023, **Sanofi**/Prometheus origin), Tulisokibart
+  (MK-7240, **Merck/MSD**), TEV-48574 (**Teva**)
+- **Competitive landscape:** Multiple companies racing — Merck/MSD, Sanofi,
+  Teva, Biocad
+- **Notes:** Emerging hot target; potential anti-fibrotic benefit unique among
+  IBD mechanisms. Multiple companies racing
+
+### Cell Therapy
+
+- **Biology:** Autologous or allogeneic cell-based therapies (MSCs, Tregs)
+- **Notes:** Mostly early-phase for perianal Crohn's disease fistulas
+
+### Microbiome / FMT
+
+- **Biology:** Restore gut microbiome via fecal microbiota transplant or defined
+  consortia
+- **Notes:** Active research area; several defined-consortium products in trials
+
+### Anti-IL-6
+
+- **Biology:** Block IL-6 or IL-6 receptor signaling
+- **Key drugs:** Olamkicept (sgp130Fc, selective IL-6 trans-signaling blocker)
+- **Notes:** Smaller class in IBD
+
+### PDE4 Inhibitor
+
+- **Biology:** Inhibit phosphodiesterase 4, reducing intracellular cAMP
+  degradation and dampening inflammatory signaling
+- **Key drugs:** Apremilast (Otezla)
+- **Notes:** Oral small molecule; primarily studied in psoriasis/psoriatic
+  arthritis, limited IBD data
+
+### Anti-IL-36
+
+- **Biology:** Block IL-36 receptor signaling, part of the IL-1 superfamily
+  involved in mucosal innate immunity
+- **Key drugs:** Spesolimab (Spevigo, Boehringer Ingelheim), Imsidolimab
+- **Notes:** Emerging target; spesolimab approved for generalized pustular
+  psoriasis, under investigation in IBD
+
+### GLP-1 / Metabolic
+
+- **Biology:** Glucagon-like peptide-1 receptor agonists and related metabolic
+  modulators with potential anti-inflammatory effects
+- **Key drugs:** Semaglutide (Ozempic/Wegovy), Tirzepatide (Mounjaro),
+  Liraglutide
+- **Notes:** Primarily metabolic agents; IBD trials exploring anti-inflammatory
+  co-benefits in obese IBD patients
+
+### Anti-p19/TL1A Bispecific
+
+- **Biology:** Bispecific antibodies simultaneously targeting IL-23 p19 and TL1A
+  for dual pathway blockade
+- **Notes:** Very early-stage; combines the two most promising emerging IBD
+  targets into a single molecule
+
+## Classification Priority
+
+The classifier checks mechanisms in this order (most specific first):
+
+1. Anti-TL1A (unique drug names, low ambiguity)
+2. Anti-IL-23 (p19) (specific drug names)
+3. Anti-IL-12/23 (p40) (ustekinumab only)
+4. Anti-TNF (well-known drugs)
+5. JAK Inhibitor
+6. Anti-Integrin
+7. S1P Modulator
+8. Anti-IL-6
+9. PDE4 Inhibitor
+10. Anti-IL-36
+11. GLP-1 / Metabolic
+12. Anti-p19/TL1A Bispecific
+13. Cell Therapy
+14. Microbiome / FMT
+15. Other Biologic (fallback for BIOLOGICAL type)
+16. Small Molecule (Other) (fallback for DRUG type)
+17. Non-pharmacological (DEVICE, BEHAVIORAL, etc.)
+18. Unclassified (no intervention info)