@kernel.chat/kbot 3.40.0 → 3.42.0

package/dist/tools/lab-bio.js

@@ -0,0 +1,1392 @@
+// kbot Life Sciences Tools — Bioinformatics, genomics, proteomics, clinical research
+// Real API integrations with NCBI, UniProt, PDB, ChEMBL, Reactome, GBIF, Open Targets, ClinicalTrials.gov
+// No external dependencies — all built on native fetch + regex XML parsing.
+import { registerTool } from './index.js';
+const UA = 'KBot/3.0 (Lab Tools)';
+// ── NCBI rate limiter (max 3 requests/sec without API key) ──────────────
+let lastNCBICall = 0;
+async function ncbiThrottle() {
+    const now = Date.now();
+    const elapsed = now - lastNCBICall;
+    if (elapsed < 334) {
+        await new Promise(resolve => setTimeout(resolve, 334 - elapsed));
+    }
+    lastNCBICall = Date.now();
+}
+// ── XML helpers (regex-based, no external parser) ───────────────────────
+function xmlTag(xml, tag) {
+    const re = new RegExp(`<${tag}[^>]*>([\\s\\S]*?)</${tag}>`, 'i');
+    const m = xml.match(re);
+    return m ? m[1].trim() : '';
+}
+function xmlTagAll(xml, tag) {
+    const re = new RegExp(`<${tag}[^>]*>([\\s\\S]*?)</${tag}>`, 'gi');
+    const results = [];
+    let m;
+    while ((m = re.exec(xml)) !== null) {
+        results.push(m[1].trim());
+    }
+    return results;
+}
+// ── Standard genetic code codon table ───────────────────────────────────
+const CODON_TABLE = {
+    TTT: 'F', TTC: 'F', TTA: 'L', TTG: 'L',
+    CTT: 'L', CTC: 'L', CTA: 'L', CTG: 'L',
+    ATT: 'I', ATC: 'I', ATA: 'I', ATG: 'M',
+    GTT: 'V', GTC: 'V', GTA: 'V', GTG: 'V',
+    TCT: 'S', TCC: 'S', TCA: 'S', TCG: 'S',
+    CCT: 'P', CCC: 'P', CCA: 'P', CCG: 'P',
+    ACT: 'T', ACC: 'T', ACA: 'T', ACG: 'T',
+    GCT: 'A', GCC: 'A', GCA: 'A', GCG: 'A',
+    TAT: 'Y', TAC: 'Y', TAA: '*', TAG: '*',
+    CAT: 'H', CAC: 'H', CAA: 'Q', CAG: 'Q',
+    AAT: 'N', AAC: 'N', AAA: 'K', AAG: 'K',
+    GAT: 'D', GAC: 'D', GAA: 'E', GAG: 'E',
+    TGT: 'C', TGC: 'C', TGA: '*', TGG: 'W',
+    CGT: 'R', CGC: 'R', CGA: 'R', CGG: 'R',
+    AGT: 'S', AGC: 'S', AGA: 'R', AGG: 'R',
+    GGT: 'G', GGC: 'G', GGA: 'G', GGG: 'G',
+};
+const AA_WEIGHTS = {
+    A: 89.09, R: 174.20, N: 132.12, D: 133.10, C: 121.16,
+    E: 147.13, Q: 146.15, G: 75.03, H: 155.16, I: 131.17,
+    L: 131.17, K: 146.19, M: 149.21, F: 165.19, P: 115.13,
+    S: 105.09, T: 119.12, W: 204.23, Y: 181.19, V: 117.15,
+};
+const DNA_WEIGHTS = {
+    A: 331.2, T: 322.2, G: 347.2, C: 307.2,
+};
+// ── Registration ────────────────────────────────────────────────────────
+export function registerLabBioTools() {
+    // ════════════════════════════════════════════════════════════════════════
+    // 1. PubMed Search
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'pubmed_search',
+        description: 'Search PubMed for biomedical literature via NCBI E-utilities. Returns titles, authors, journal, year, abstract, DOI, and PMID. Use MeSH terms for precise filtering.',
+        parameters: {
+            query: { type: 'string', description: 'Search query (e.g., "CRISPR cancer therapy")', required: true },
+            mesh_terms: { type: 'string', description: 'Optional MeSH terms to AND with query (e.g., "Neoplasms[MeSH]")' },
+            limit: { type: 'number', description: 'Max results (default 10, max 50)' },
+            sort: { type: 'string', description: 'Sort order: "relevance" (default) or "date"' },
+        },
+        tier: 'free',
+        async execute(args) {
+            const query = String(args.query);
+            const meshTerms = args.mesh_terms ? String(args.mesh_terms) : '';
+            const limit = Math.min(typeof args.limit === 'number' ? args.limit : 10, 50);
+            const sort = String(args.sort || 'relevance');
+            let fullQuery = query;
+            if (meshTerms)
+                fullQuery += ` AND ${meshTerms}`;
+            const sortParam = sort === 'date' ? '&sort=pub_date' : '&sort=relevance';
+            try {
+                // Step 1: esearch to get PMIDs
+                await ncbiThrottle();
+                const searchUrl = `https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&retmode=json&retmax=${limit}${sortParam}&term=${encodeURIComponent(fullQuery)}`;
+                const searchRes = await fetch(searchUrl, {
+                    headers: { 'User-Agent': UA },
+                    signal: AbortSignal.timeout(10000),
+                });
+                if (!searchRes.ok)
+                    return `PubMed search error: HTTP ${searchRes.status}`;
+                const searchData = await searchRes.json();
+                const idList = searchData?.esearchresult?.idlist || [];
+                const totalCount = searchData?.esearchresult?.count || '0';
+                if (idList.length === 0)
+                    return `No PubMed results for "${query}". Try broader terms or check MeSH vocabulary.`;
+                // Step 2: efetch to get article details
+                await ncbiThrottle();
+                const fetchUrl = `https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&retmode=xml&id=${idList.join(',')}`;
+                const fetchRes = await fetch(fetchUrl, {
+                    headers: { 'User-Agent': UA },
+                    signal: AbortSignal.timeout(10000),
+                });
+                if (!fetchRes.ok)
+                    return `PubMed fetch error: HTTP ${fetchRes.status}`;
+                const xml = await fetchRes.text();
+                // Parse articles from XML
+                const articles = xmlTagAll(xml, 'PubmedArticle');
+                const results = [`## PubMed Results (${totalCount} total, showing ${articles.length})\n`];
+                for (const article of articles) {
+                    const pmid = xmlTag(article, 'PMID');
+                    const title = xmlTag(article, 'ArticleTitle').replace(/<[^>]+>/g, '');
+                    const abstractText = xmlTag(article, 'AbstractText').replace(/<[^>]+>/g, '');
+                    const journal = xmlTag(article, 'Title');
+                    const year = xmlTag(article, 'Year');
+                    // Authors
+                    const authorNodes = xmlTagAll(article, 'Author');
+                    const authors = authorNodes.slice(0, 5).map(a => {
+                        const last = xmlTag(a, 'LastName');
+                        const initials = xmlTag(a, 'Initials');
+                        return last ? `${last} ${initials}` : xmlTag(a, 'CollectiveName');
+                    }).filter(Boolean);
+                    const authorStr = authors.join(', ') + (authorNodes.length > 5 ? ' et al.' : '');
+                    // DOI
+                    const articleIdList = xmlTagAll(article, 'ArticleId');
+                    let doi = '';
+                    for (const idBlock of articleIdList) {
+                        if (idBlock.includes('doi')) {
+                            // The DOI is the text content, but we need to check the IdType attribute
+                            const doiMatch = article.match(/<ArticleId IdType="doi">([^<]+)<\/ArticleId>/i);
+                            if (doiMatch)
+                                doi = doiMatch[1];
+                            break;
+                        }
+                    }
+                    let entry = `### ${title}\n`;
+                    entry += `**Authors:** ${authorStr || 'N/A'}\n`;
+                    entry += `**Journal:** ${journal || 'N/A'} (${year || 'N/A'})\n`;
+                    entry += `**PMID:** [${pmid}](https://pubmed.ncbi.nlm.nih.gov/${pmid}/)`;
+                    if (doi)
+                        entry += ` | **DOI:** [${doi}](https://doi.org/${doi})`;
+                    entry += '\n';
+                    if (abstractText) {
+                        const truncated = abstractText.length > 500 ? abstractText.slice(0, 500) + '...' : abstractText;
+                        entry += `\n> ${truncated}\n`;
+                    }
+                    results.push(entry);
+                }
+                return results.join('\n---\n');
+            }
+            catch (e) {
+                return `PubMed search failed: ${e?.message || e}`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 2. Gene Lookup
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'gene_lookup',
+        description: 'Look up gene information by symbol or name using MyGene.info. Returns function, chromosome location, aliases, and associated diseases.',
+        parameters: {
+            gene: { type: 'string', description: 'Gene symbol or name (e.g., "TP53", "BRCA1", "tumor protein p53")', required: true },
+            organism: { type: 'string', description: 'Organism (default: "human"). Also: "mouse", "rat", "zebrafish", etc.' },
+        },
+        tier: 'free',
+        async execute(args) {
+            const gene = String(args.gene);
+            const organism = String(args.organism || 'human');
+            try {
+                const fields = 'symbol,name,summary,genomic_pos,alias,type_of_gene,entrezgene,ensembl.gene,taxid,generif,pathway.kegg';
+                const url = `https://mygene.info/v3/query?q=${encodeURIComponent(gene)}&species=${encodeURIComponent(organism)}&fields=${fields}&size=5`;
+                const res = await fetch(url, {
+                    headers: { 'User-Agent': UA },
+                    signal: AbortSignal.timeout(10000),
+                });
+                if (!res.ok)
+                    return `MyGene.info error: HTTP ${res.status}`;
+                const data = await res.json();
+                const hits = data.hits || [];
+                if (hits.length === 0)
+                    return `No gene found for "${gene}" in ${organism}. Try the official HGNC symbol.`;
+                const results = [`## Gene Lookup: "${gene}" (${organism})\n`];
+                for (const hit of hits.slice(0, 3)) {
+                    let entry = `### ${hit.symbol || gene} — ${hit.name || 'Unknown'}\n`;
+                    entry += `**Type:** ${hit.type_of_gene || 'N/A'}\n`;
+                    if (hit.entrezgene)
+                        entry += `**Entrez ID:** [${hit.entrezgene}](https://www.ncbi.nlm.nih.gov/gene/${hit.entrezgene})\n`;
+                    if (hit.ensembl?.gene)
+                        entry += `**Ensembl:** ${hit.ensembl.gene}\n`;
+                    // Genomic position
+                    if (hit.genomic_pos) {
+                        const pos = Array.isArray(hit.genomic_pos) ? hit.genomic_pos[0] : hit.genomic_pos;
+                        if (pos)
+                            entry += `**Location:** Chr${pos.chr}:${pos.start?.toLocaleString()}-${pos.end?.toLocaleString()} (${pos.strand > 0 ? '+' : '-'} strand)\n`;
+                    }
+                    // Aliases
+                    if (hit.alias) {
+                        const aliases = Array.isArray(hit.alias) ? hit.alias : [hit.alias];
+                        entry += `**Aliases:** ${aliases.slice(0, 10).join(', ')}\n`;
+                    }
+                    // Summary
+                    if (hit.summary) {
+                        const truncated = hit.summary.length > 600 ? hit.summary.slice(0, 600) + '...' : hit.summary;
+                        entry += `\n**Summary:** ${truncated}\n`;
+                    }
+                    // Pathways
+                    if (hit.pathway?.kegg) {
+                        const pathways = Array.isArray(hit.pathway.kegg) ? hit.pathway.kegg : [hit.pathway.kegg];
+                        entry += `\n**KEGG Pathways:**\n`;
+                        for (const p of pathways.slice(0, 5)) {
+                            entry += `- ${p.name || p.id}\n`;
+                        }
+                    }
+                    results.push(entry);
+                }
+                return results.join('\n---\n');
+            }
+            catch (e) {
+                return `Gene lookup failed: ${e?.message || e}`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 3. Protein Search
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'protein_search',
+        description: 'Search UniProt for protein information. Returns sequence length, function, subcellular location, and GO terms.',
+        parameters: {
+            query: { type: 'string', description: 'Protein name, gene symbol, or keyword (e.g., "insulin", "P53_HUMAN")', required: true },
+            organism: { type: 'string', description: 'Organism filter (e.g., "Homo sapiens", "Mus musculus")' },
+            reviewed: { type: 'boolean', description: 'Only reviewed (Swiss-Prot) entries (default: true)' },
+        },
+        tier: 'free',
+        async execute(args) {
+            const query = String(args.query);
+            const organism = args.organism ? String(args.organism) : '';
+            const reviewed = args.reviewed !== false;
+            try {
+                let fullQuery = query;
+                if (organism)
+                    fullQuery += ` AND organism_name:"${organism}"`;
+                if (reviewed)
+                    fullQuery += ' AND reviewed:true';
+                const url = `https://rest.uniprot.org/uniprotkb/search?query=${encodeURIComponent(fullQuery)}&format=json&size=5`;
+                const res = await fetch(url, {
+                    headers: { 'User-Agent': UA },
+                    signal: AbortSignal.timeout(10000),
+                });
+                if (!res.ok)
+                    return `UniProt error: HTTP ${res.status}`;
+                const data = await res.json();
+                const results_arr = data.results || [];
+                if (results_arr.length === 0)
+                    return `No UniProt results for "${query}". Try a broader query or set reviewed=false.`;
+                const output = [`## UniProt Search: "${query}"\n`];
+                for (const entry of results_arr) {
+                    const accession = entry.primaryAccession || 'N/A';
+                    const entryName = entry.uniProtkbId || '';
+                    const proteinName = entry.proteinDescription?.recommendedName?.fullName?.value
+                        || entry.proteinDescription?.submittedName?.[0]?.fullName?.value
+                        || 'Unknown';
+                    const organism_name = entry.organism?.scientificName || '';
+                    const seqLen = entry.sequence?.length || 0;
+                    const seqMW = entry.sequence?.molWeight || 0;
+                    let block = `### ${proteinName}\n`;
+                    block += `**Accession:** [${accession}](https://www.uniprot.org/uniprot/${accession}) (${entryName})\n`;
+                    block += `**Organism:** ${organism_name}\n`;
+                    block += `**Sequence:** ${seqLen} aa | ${(seqMW / 1000).toFixed(1)} kDa\n`;
+                    // Function
+                    const funcComments = (entry.comments || []).filter((c) => c.commentType === 'FUNCTION');
+                    if (funcComments.length > 0) {
+                        const funcText = funcComments[0].texts?.[0]?.value || '';
+                        if (funcText) {
+                            const truncated = funcText.length > 400 ? funcText.slice(0, 400) + '...' : funcText;
+                            block += `\n**Function:** ${truncated}\n`;
+                        }
+                    }
+                    // Subcellular location
+                    const locComments = (entry.comments || []).filter((c) => c.commentType === 'SUBCELLULAR LOCATION');
+                    if (locComments.length > 0) {
+                        const locations = locComments[0].subcellularLocations?.map((sl) => sl.location?.value).filter(Boolean) || [];
+                        if (locations.length > 0)
+                            block += `**Subcellular Location:** ${locations.join(', ')}\n`;
+                    }
+                    // GO terms
+                    const goTerms = (entry.uniProtKBCrossReferences || []).filter((x) => x.database === 'GO');
+                    if (goTerms.length > 0) {
+                        const goGroups = { F: [], P: [], C: [] };
+                        for (const go of goTerms.slice(0, 20)) {
+                            const name = go.properties?.find((p) => p.key === 'GoTerm')?.value || go.id;
+                            if (name.startsWith('F:'))
+                                goGroups.F.push(name.slice(2));
+                            else if (name.startsWith('P:'))
+                                goGroups.P.push(name.slice(2));
+                            else if (name.startsWith('C:'))
+                                goGroups.C.push(name.slice(2));
+                        }
+                        if (goGroups.F.length > 0)
+                            block += `**Molecular Function:** ${goGroups.F.slice(0, 5).join('; ')}\n`;
+                        if (goGroups.P.length > 0)
+                            block += `**Biological Process:** ${goGroups.P.slice(0, 5).join('; ')}\n`;
+                        if (goGroups.C.length > 0)
+                            block += `**Cellular Component:** ${goGroups.C.slice(0, 5).join('; ')}\n`;
+                    }
+                    output.push(block);
+                }
+                return output.join('\n---\n');
+            }
+            catch (e) {
+                return `Protein search failed: ${e?.message || e}`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 4. Protein Structure (PDB)
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'protein_structure',
+        description: 'Fetch protein 3D structure info from RCSB PDB. Get resolution, experimental method, ligands, and chain details by PDB ID or text search.',
+        parameters: {
+            pdb_id: { type: 'string', description: 'PDB ID (e.g., "1TUP", "6LU7"). If provided, fetches directly.' },
+            query: { type: 'string', description: 'Text search query (e.g., "p53 DNA binding domain"). Used if pdb_id not given.' },
+        },
+        tier: 'free',
+        async execute(args) {
+            const pdbId = args.pdb_id ? String(args.pdb_id).toUpperCase() : '';
+            const query = args.query ? String(args.query) : '';
+            if (!pdbId && !query)
+                return 'Provide either pdb_id or query to search PDB structures.';
+            try {
+                let ids = [];
+                if (pdbId) {
+                    ids = [pdbId];
+                }
+                else {
+                    // Text search via RCSB search API
+                    const searchBody = JSON.stringify({
+                        query: {
+                            type: 'terminal',
+                            service: 'full_text',
+                            parameters: { value: query },
+                        },
+                        return_type: 'entry',
+                        request_options: { paginate: { start: 0, rows: 5 } },
+                    });
+                    const searchRes = await fetch('https://search.rcsb.org/rcsbsearch/v2/query', {
+                        method: 'POST',
+                        headers: { 'Content-Type': 'application/json', 'User-Agent': UA },
+                        body: searchBody,
+                        signal: AbortSignal.timeout(10000),
+                    });
+                    if (!searchRes.ok)
+                        return `PDB search error: HTTP ${searchRes.status}`;
+                    const searchData = await searchRes.json();
+                    ids = (searchData.result_set || []).map((r) => r.identifier).slice(0, 5);
+                    if (ids.length === 0)
+                        return `No PDB structures found for "${query}".`;
+                }
+                const output = [`## PDB Structure${ids.length > 1 ? 's' : ''}\n`];
+                for (const id of ids) {
+                    const res = await fetch(`https://data.rcsb.org/rest/v1/core/entry/${id}`, {
+                        headers: { 'User-Agent': UA },
+                        signal: AbortSignal.timeout(10000),
+                    });
+                    if (!res.ok) {
+                        output.push(`**${id}**: Not found (HTTP ${res.status})`);
+                        continue;
+                    }
+                    const data = await res.json();
+                    const title = data.struct?.title || 'No title';
+                    const method = data.exptl?.[0]?.method || 'N/A';
+                    const resolution = data.rcsb_entry_info?.resolution_combined?.[0];
+                    const deposited = data.rcsb_accession_info?.deposit_date || '';
+                    const polymerCount = data.rcsb_entry_info?.polymer_entity_count || 0;
+                    const nonPolymerCount = data.rcsb_entry_info?.nonpolymer_entity_count || 0;
+                    const citation = data.rcsb_primary_citation;
+                    let block = `### [${id}](https://www.rcsb.org/structure/${id}) — ${title}\n`;
+                    block += `**Method:** ${method}`;
+                    if (resolution)
+                        block += ` | **Resolution:** ${resolution} A`;
+                    block += '\n';
+                    block += `**Deposited:** ${deposited}\n`;
+                    block += `**Entities:** ${polymerCount} polymer, ${nonPolymerCount} non-polymer (ligands/ions)\n`;
+                    // Polymer entities (chains)
+                    if (data.rcsb_entry_info?.polymer_entity_count_protein) {
+                        block += `**Protein chains:** ${data.rcsb_entry_info.polymer_entity_count_protein}\n`;
+                    }
+                    if (data.rcsb_entry_info?.polymer_entity_count_nucleic_acid) {
+                        block += `**Nucleic acid chains:** ${data.rcsb_entry_info.polymer_entity_count_nucleic_acid}\n`;
+                    }
+                    // Primary citation
+                    if (citation) {
+                        block += `\n**Citation:** ${citation.title || ''}\n`;
+                        const authors = citation.rcsb_authors?.slice(0, 3).join(', ') || '';
+                        if (authors)
+                            block += `*${authors}${citation.rcsb_authors?.length > 3 ? ' et al.' : ''}* `;
+                        if (citation.pdbx_database_id_journal)
+                            block += `${citation.pdbx_database_id_journal} `;
+                        if (citation.year)
+                            block += `(${citation.year})`;
+                        if (citation.pdbx_database_id_DOI)
+                            block += ` DOI: ${citation.pdbx_database_id_DOI}`;
+                        block += '\n';
+                    }
+                    output.push(block);
+                }
+                return output.join('\n---\n');
+            }
+            catch (e) {
+                return `PDB lookup failed: ${e?.message || e}`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 5. BLAST Search
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'blast_search',
+        description: 'Submit a sequence for NCBI BLAST homology search. Supports blastn (nucleotide), blastp (protein), blastx (translated). Async — submits job, polls for results.',
+        parameters: {
+            sequence: { type: 'string', description: 'Nucleotide or protein sequence (FASTA or raw)', required: true },
+            program: { type: 'string', description: 'BLAST program: "blastn", "blastp", or "blastx"', required: true },
+            database: { type: 'string', description: 'Database: "nr" (non-redundant protein), "nt" (nucleotide), "swissprot"', required: true },
+        },
+        tier: 'free',
+        timeout: 180_000,
+        async execute(args) {
+            const sequence = String(args.sequence).trim();
+            const program = String(args.program || 'blastn');
+            const database = String(args.database || 'nr');
+            if (sequence.length < 10)
+                return 'Sequence too short for BLAST. Provide at least 10 residues/bases.';
+            // Clean sequence: remove FASTA header if present
+            const cleanSeq = sequence.split('\n').filter(line => !line.startsWith('>')).join('');
+            try {
+                // Step 1: Submit BLAST job
+                await ncbiThrottle();
+                const submitParams = new URLSearchParams({
+                    CMD: 'Put',
+                    PROGRAM: program,
+                    DATABASE: database,
+                    QUERY: cleanSeq,
+                    FORMAT_TYPE: 'XML',
+                    HITLIST_SIZE: '10',
+                });
+                const submitRes = await fetch('https://blast.ncbi.nlm.nih.gov/blast/Blast.cgi', {
+                    method: 'POST',
+                    headers: { 'User-Agent': UA, 'Content-Type': 'application/x-www-form-urlencoded' },
+                    body: submitParams.toString(),
+                    signal: AbortSignal.timeout(30000),
+                });
+                if (!submitRes.ok)
+                    return `BLAST submission failed: HTTP ${submitRes.status}`;
+                const submitText = await submitRes.text();
+                // Extract RID from response
+                const ridMatch = submitText.match(/RID\s*=\s*(\S+)/);
+                if (!ridMatch)
+                    return `BLAST submission failed: could not get Request ID.\n\nResponse excerpt:\n${submitText.slice(0, 500)}`;
+                const rid = ridMatch[1];
+                // Extract estimated wait time
+                const rtoeMatch = submitText.match(/RTOE\s*=\s*(\d+)/);
+                const rtoe = rtoeMatch ? parseInt(rtoeMatch[1], 10) : 15;
+                // Step 2: Poll for results
+                const startTime = Date.now();
+                const maxWait = 120_000; // 2 minutes max polling
+                let waitTime = Math.min(rtoe * 1000, 15000); // Start with estimated wait, cap at 15s
+                // Initial wait before first poll
+                await new Promise(resolve => setTimeout(resolve, Math.min(waitTime, 10000)));
+                let resultXml = '';
+                while (Date.now() - startTime < maxWait) {
+                    await ncbiThrottle();
+                    const pollUrl = `https://blast.ncbi.nlm.nih.gov/blast/Blast.cgi?CMD=Get&FORMAT_TYPE=XML&RID=${rid}`;
+                    const pollRes = await fetch(pollUrl, {
+                        headers: { 'User-Agent': UA },
+                        signal: AbortSignal.timeout(120_000),
+                    });
+                    if (!pollRes.ok) {
+                        await new Promise(resolve => setTimeout(resolve, 5000));
+                        continue;
+                    }
+                    const pollText = await pollRes.text();
+                    // Check status
+                    if (pollText.includes('Status=WAITING')) {
+                        await new Promise(resolve => setTimeout(resolve, 5000));
+                        continue;
+                    }
+                    if (pollText.includes('Status=FAILED'))
+                        return `BLAST job failed (RID: ${rid}). The sequence may be invalid for ${program}.`;
+                    if (pollText.includes('Status=UNKNOWN'))
+                        return `BLAST job expired or unknown (RID: ${rid}).`;
+                    // If we have actual results (XML with hits)
+                    if (pollText.includes('<BlastOutput>') || pollText.includes('<Hit>')) {
+                        resultXml = pollText;
+                        break;
+                    }
+                    // Still processing
+                    await new Promise(resolve => setTimeout(resolve, 5000));
+                }
+                if (!resultXml)
+                    return `BLAST search timed out after ${Math.round(maxWait / 1000)}s. RID: ${rid} — check manually at https://blast.ncbi.nlm.nih.gov/blast/Blast.cgi?CMD=Get&RID=${rid}`;
+                // Step 3: Parse results
+                const hits = xmlTagAll(resultXml, 'Hit');
+                if (hits.length === 0)
+                    return `BLAST completed but found no significant hits for your ${program} search against ${database}.`;
+                const output = [`## BLAST Results (${program} vs ${database})\n**RID:** ${rid} | **Hits:** ${hits.length}\n`];
+                for (const hit of hits.slice(0, 10)) {
+                    const hitNum = xmlTag(hit, 'Hit_num');
+                    const hitDef = xmlTag(hit, 'Hit_def').slice(0, 120);
+                    const hitAccession = xmlTag(hit, 'Hit_accession');
+                    const hitLen = xmlTag(hit, 'Hit_len');
+                    // Best HSP
+                    const hsps = xmlTagAll(hit, 'Hsp');
+                    const hsp = hsps[0] || '';
+                    const evalue = xmlTag(hsp, 'Hsp_evalue');
+                    const bitScore = xmlTag(hsp, 'Hsp_bit-score');
+                    const identity = xmlTag(hsp, 'Hsp_identity');
+                    const alignLen = xmlTag(hsp, 'Hsp_align-len');
+                    const identPct = alignLen ? ((parseInt(identity, 10) / parseInt(alignLen, 10)) * 100).toFixed(1) : 'N/A';
+                    let block = `**${hitNum}. ${hitDef}**\n`;
+                    block += `Accession: ${hitAccession} | Length: ${hitLen}\n`;
+                    block += `E-value: ${evalue} | Bit score: ${bitScore} | Identity: ${identPct}% (${identity}/${alignLen})\n`;
+                    output.push(block);
+                }
+                return output.join('\n---\n');
+            }
+            catch (e) {
+                return `BLAST search failed: ${e?.message || e}`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 6. Drug Lookup (ChEMBL)
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'drug_lookup',
+        description: 'Search ChEMBL for drugs and compounds. Returns targets, mechanism of action, clinical phase, and molecular properties.',
+        parameters: {
+            query: { type: 'string', description: 'Drug name, target, or mechanism (e.g., "imatinib", "EGFR inhibitor")', required: true },
+            search_type: { type: 'string', description: 'Search type: "name" (default), "target", or "mechanism"' },
+        },
+        tier: 'free',
+        async execute(args) {
+            const query = String(args.query);
+            const searchType = String(args.search_type || 'name');
+            try {
+                let url;
+                if (searchType === 'target') {
+                    url = `https://www.ebi.ac.uk/chembl/api/data/target/search?q=${encodeURIComponent(query)}&format=json&limit=10`;
+                }
+                else if (searchType === 'mechanism') {
+                    url = `https://www.ebi.ac.uk/chembl/api/data/mechanism/search?q=${encodeURIComponent(query)}&format=json&limit=10`;
+                }
+                else {
+                    url = `https://www.ebi.ac.uk/chembl/api/data/molecule/search?q=${encodeURIComponent(query)}&format=json&limit=10`;
+                }
+                const res = await fetch(url, {
+                    headers: { 'User-Agent': UA },
+                    signal: AbortSignal.timeout(10000),
+                });
+                if (!res.ok)
+                    return `ChEMBL error: HTTP ${res.status}`;
+                const data = await res.json();
+                if (searchType === 'target') {
+                    const targets = data.targets || [];
+                    if (targets.length === 0)
+                        return `No targets found for "${query}" in ChEMBL.`;
+                    const output = [`## ChEMBL Targets for "${query}"\n`];
+                    for (const t of targets.slice(0, 5)) {
+                        let block = `### ${t.pref_name || 'Unknown'}\n`;
+                        block += `**ChEMBL ID:** ${t.target_chembl_id || 'N/A'}\n`;
+                        block += `**Type:** ${t.target_type || 'N/A'}\n`;
+                        block += `**Organism:** ${t.organism || 'N/A'}\n`;
+                        if (t.target_components?.[0]?.accession) {
+                            block += `**UniProt:** ${t.target_components[0].accession}\n`;
+                        }
+                        output.push(block);
+                    }
+                    return output.join('\n---\n');
+                }
+                if (searchType === 'mechanism') {
+                    const mechanisms = data.mechanisms || [];
+                    if (mechanisms.length === 0)
+                        return `No mechanisms found for "${query}" in ChEMBL.`;
+                    const output = [`## ChEMBL Mechanisms: "${query}"\n`];
+                    for (const m of mechanisms.slice(0, 10)) {
+                        let block = `**${m.molecule_chembl_id}** → ${m.target_chembl_id || 'N/A'}\n`;
+                        block += `Mechanism: ${m.mechanism_of_action || 'N/A'}\n`;
+                        block += `Action type: ${m.action_type || 'N/A'}\n`;
+                        if (m.max_phase !== undefined)
+                            block += `Max phase: ${m.max_phase}\n`;
+                        output.push(block);
+                    }
+                    return output.join('\n---\n');
+                }
+                // Default: molecule search
+                const molecules = data.molecules || [];
+                if (molecules.length === 0)
+                    return `No molecules found for "${query}" in ChEMBL.`;
+                const output = [`## ChEMBL Molecules: "${query}"\n`];
+                for (const mol of molecules.slice(0, 5)) {
+                    const name = mol.pref_name || mol.molecule_chembl_id || 'Unknown';
+                    const chemblId = mol.molecule_chembl_id || 'N/A';
+                    const maxPhase = mol.max_phase !== undefined ? mol.max_phase : 'N/A';
+                    const type = mol.molecule_type || 'N/A';
+                    const props = mol.molecule_properties || {};
+                    let block = `### ${name}\n`;
+                    block += `**ChEMBL ID:** [${chemblId}](https://www.ebi.ac.uk/chembl/compound_report_card/${chemblId}/)\n`;
+                    block += `**Type:** ${type} | **Max Clinical Phase:** ${maxPhase}\n`;
+                    if (mol.first_approval)
+                        block += `**First Approval:** ${mol.first_approval}\n`;
+                    // Molecular properties
+                    if (props.full_mwt)
+                        block += `**MW:** ${props.full_mwt} Da`;
+                    if (props.alogp)
+                        block += ` | **ALogP:** ${props.alogp}`;
+                    if (props.hba)
+                        block += ` | **HBA:** ${props.hba}`;
+                    if (props.hbd)
+                        block += ` | **HBD:** ${props.hbd}`;
+                    if (props.psa)
+                        block += ` | **PSA:** ${props.psa}`;
+                    if (props.full_mwt)
+                        block += '\n';
+                    if (props.ro5_violations !== undefined)
+                        block += `**Lipinski violations:** ${props.ro5_violations}\n`;
+                    if (mol.molecule_structures?.canonical_smiles) {
+                        block += `**SMILES:** \`${mol.molecule_structures.canonical_smiles.slice(0, 100)}\`\n`;
+                    }
+                    output.push(block);
+                }
+                return output.join('\n---\n');
+            }
+            catch (e) {
+                return `Drug lookup failed: ${e?.message || e}`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 7. Pathway Search (Reactome)
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'pathway_search',
+        description: 'Search Reactome for biological pathways. Returns pathway names, species, summaries, and hierarchical relationships.',
+        parameters: {
+            query: { type: 'string', description: 'Pathway name or keyword (e.g., "apoptosis", "glycolysis", "MAPK signaling")', required: true },
+            organism: { type: 'string', description: 'Species name (default: "Homo sapiens")' },
+        },
+        tier: 'free',
+        async execute(args) {
+            const query = String(args.query);
+            const organism = String(args.organism || 'Homo sapiens');
+            try {
+                const url = `https://reactome.org/ContentService/search/query?query=${encodeURIComponent(query)}&species=${encodeURIComponent(organism)}&types=Pathway&cluster=true`;
+                const res = await fetch(url, {
+                    headers: { 'User-Agent': UA, 'Accept': 'application/json' },
+                    signal: AbortSignal.timeout(10000),
+                });
+                if (!res.ok)
+                    return `Reactome error: HTTP ${res.status}`;
+                const data = await res.json();
+                const groups = data.results || [];
+                if (groups.length === 0)
+                    return `No pathways found for "${query}" in Reactome.`;
+                const output = [`## Reactome Pathways: "${query}"\n`];
+                let count = 0;
+                for (const group of groups) {
+                    const entries = group.entries || [];
+                    for (const entry of entries) {
+                        if (count >= 10)
+                            break;
+                        let block = `### ${entry.name || 'Unknown'}\n`;
+                        block += `**ID:** [${entry.stId}](https://reactome.org/content/detail/${entry.stId})\n`;
+                        block += `**Species:** ${entry.species?.[0] || organism}\n`;
+                        if (entry.summation) {
+                            const summary = entry.summation.replace(/<[^>]+>/g, '');
+                            const truncated = summary.length > 300 ? summary.slice(0, 300) + '...' : summary;
+                            block += `\n${truncated}\n`;
+                        }
+                        if (entry.compartmentNames?.length > 0) {
+                            block += `**Compartments:** ${entry.compartmentNames.join(', ')}\n`;
+                        }
+                        output.push(block);
+                        count++;
+                    }
+                    if (count >= 10)
+                        break;
+                }
+                return output.join('\n---\n');
+            }
+            catch (e) {
+                return `Pathway search failed: ${e?.message || e}`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 8. Taxonomy Lookup (GBIF)
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'taxonomy_lookup',
+        description: 'Look up taxonomic classification of any organism via GBIF. Returns full lineage from kingdom to species with taxonomic status.',
+        parameters: {
+            name: { type: 'string', description: 'Organism name (e.g., "Homo sapiens", "E. coli", "giant panda")', required: true },
+            rank: { type: 'string', description: 'Expected rank filter: "species", "genus", "family", etc.' },
+        },
+        tier: 'free',
+        async execute(args) {
+            const name = String(args.name);
+            const rank = args.rank ? String(args.rank).toUpperCase() : '';
+            try {
+                let url = `https://api.gbif.org/v1/species/search?q=${encodeURIComponent(name)}&limit=5`;
+                if (rank)
+                    url += `&rank=${rank}`;
+                const res = await fetch(url, {
+                    headers: { 'User-Agent': UA },
+                    signal: AbortSignal.timeout(10000),
+                });
+                if (!res.ok)
+                    return `GBIF error: HTTP ${res.status}`;
+                const data = await res.json();
+                const results_arr = data.results || [];
+                if (results_arr.length === 0)
+                    return `No taxonomic records for "${name}" in GBIF.`;
+                const output = [`## Taxonomy: "${name}"\n`];
+                for (const sp of results_arr.slice(0, 3)) {
+                    let block = `### ${sp.canonicalName || sp.scientificName || name}\n`;
+                    if (sp.authorship)
+                        block += `*${sp.authorship}*\n`;
+                    block += `**Rank:** ${sp.rank || 'N/A'}\n`;
+                    block += `**Status:** ${sp.taxonomicStatus || 'N/A'}\n`;
+                    block += `**GBIF Key:** [${sp.key}](https://www.gbif.org/species/${sp.key})\n`;
+                    // Full lineage
+                    const lineage = [];
+                    if (sp.kingdom)
+                        lineage.push(`Kingdom: ${sp.kingdom}`);
+                    if (sp.phylum)
+                        lineage.push(`Phylum: ${sp.phylum}`);
+                    if (sp.class)
+                        lineage.push(`Class: ${sp.class}`);
+                    if (sp.order)
+                        lineage.push(`Order: ${sp.order}`);
+                    if (sp.family)
+                        lineage.push(`Family: ${sp.family}`);
+                    if (sp.genus)
+                        lineage.push(`Genus: *${sp.genus}*`);
+                    if (sp.species)
+                        lineage.push(`Species: *${sp.species}*`);
+                    if (lineage.length > 0) {
+                        block += `\n**Lineage:**\n${lineage.map(l => `- ${l}`).join('\n')}\n`;
+                    }
+                    if (sp.vernacularNames?.length > 0) {
+                        const common = sp.vernacularNames.slice(0, 5).map((v) => v.vernacularName).filter(Boolean);
+                        if (common.length > 0)
+                            block += `\n**Common Names:** ${common.join(', ')}\n`;
+                    }
+                    if (sp.descriptions?.length > 0) {
+                        const desc = sp.descriptions[0].description || '';
+                        if (desc)
+                            block += `\n${desc.slice(0, 300)}${desc.length > 300 ? '...' : ''}\n`;
+                    }
+                    output.push(block);
+                }
+                return output.join('\n---\n');
+            }
+            catch (e) {
+                return `Taxonomy lookup failed: ${e?.message || e}`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 9. Clinical Trials
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'clinical_trials',
+        description: 'Search ClinicalTrials.gov for clinical studies. Filter by condition, drug, status, and phase.',
+        parameters: {
+            query: { type: 'string', description: 'Search term (e.g., "pembrolizumab melanoma", "COVID-19 vaccine")', required: true },
+            status: { type: 'string', description: 'Filter: "recruiting", "completed", "active" (active, not recruiting), "enrolling" (enrolling by invitation)' },
+            phase: { type: 'string', description: 'Phase filter: "EARLY_PHASE1", "PHASE1", "PHASE2", "PHASE3", "PHASE4"' },
+        },
+        tier: 'free',
+        async execute(args) {
+            const query = String(args.query);
+            const status = args.status ? String(args.status) : '';
+            const phase = args.phase ? String(args.phase) : '';
+            try {
+                let url = `https://clinicaltrials.gov/api/v2/studies?query.term=${encodeURIComponent(query)}&pageSize=10`;
+                // Map user-friendly status to API values
+                if (status) {
+                    const statusMap = {
+                        recruiting: 'RECRUITING',
+                        completed: 'COMPLETED',
+                        active: 'ACTIVE_NOT_RECRUITING',
+                        enrolling: 'ENROLLING_BY_INVITATION',
+                    };
+                    const mapped = statusMap[status.toLowerCase()] || status.toUpperCase();
+                    url += `&filter.overallStatus=${mapped}`;
+                }
+                if (phase) {
+                    url += `&filter.phase=${phase.toUpperCase()}`;
+                }
+                const res = await fetch(url, {
+                    headers: { 'User-Agent': UA },
+                    signal: AbortSignal.timeout(10000),
+                });
+                if (!res.ok)
+                    return `ClinicalTrials.gov error: HTTP ${res.status}`;
+                const data = await res.json();
+                const studies = data.studies || [];
+                if (studies.length === 0)
+                    return `No clinical trials found for "${query}".`;
+                const totalCount = data.totalCount || studies.length;
+                const output = [`## Clinical Trials: "${query}" (${totalCount} total)\n`];
+                for (const study of studies) {
+                    const proto = study.protocolSection || {};
+                    const id_module = proto.identificationModule || {};
+                    const status_module = proto.statusModule || {};
+                    const design_module = proto.designModule || {};
+                    const desc_module = proto.descriptionModule || {};
+                    const conditions_module = proto.conditionsModule || {};
+                    const sponsor_module = proto.sponsorCollaboratorsModule || {};
+                    const nctId = id_module.nctId || 'N/A';
+                    const title = id_module.officialTitle || id_module.briefTitle || 'Untitled';
+                    const overallStatus = status_module.overallStatus || 'N/A';
+                    const phases = design_module.phases?.join(', ') || 'N/A';
+                    const startDate = status_module.startDateStruct?.date || '';
+                    const completionDate = status_module.completionDateStruct?.date || '';
+                    const briefSummary = desc_module.briefSummary || '';
+                    const conditions = conditions_module.conditions?.join(', ') || '';
+                    const sponsor = sponsor_module.leadSponsor?.name || '';
+                    let block = `### ${title.slice(0, 150)}\n`;
+                    block += `**NCT ID:** [${nctId}](https://clinicaltrials.gov/study/${nctId})\n`;
+                    block += `**Status:** ${overallStatus} | **Phase:** ${phases}\n`;
+                    if (sponsor)
+                        block += `**Sponsor:** ${sponsor}\n`;
+                    if (conditions)
+                        block += `**Conditions:** ${conditions}\n`;
+                    if (startDate)
+                        block += `**Dates:** ${startDate}${completionDate ? ` → ${completionDate}` : ''}\n`;
+                    if (briefSummary) {
+                        const truncated = briefSummary.length > 300 ? briefSummary.slice(0, 300) + '...' : briefSummary;
+                        block += `\n> ${truncated}\n`;
+                    }
+                    output.push(block);
+                }
+                return output.join('\n---\n');
+            }
+            catch (e) {
+                return `Clinical trials search failed: ${e?.message || e}`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 10. Disease Info (Open Targets)
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'disease_info',
+        description: 'Look up disease information from Open Targets Platform. Returns associated genes, drugs, and therapeutic areas via GraphQL.',
+        parameters: {
+            disease: { type: 'string', description: 'Disease name (e.g., "lung cancer", "Alzheimer", "diabetes mellitus")', required: true },
+        },
+        tier: 'free',
+        async execute(args) {
+            const disease = String(args.disease);
+            try {
+                // Step 1: Search for disease ID
+                const searchQuery = `
+          query SearchDisease($q: String!) {
+            search(queryString: $q, entityNames: ["disease"], page: { index: 0, size: 3 }) {
+              hits {
+                id
+                entity
+                name
+                description
+              }
+              total
+            }
+          }
+        `;
+                const searchRes = await fetch('https://api.platform.opentargets.org/api/v4/graphql', {
+                    method: 'POST',
+                    headers: { 'Content-Type': 'application/json', 'User-Agent': UA },
+                    body: JSON.stringify({ query: searchQuery, variables: { q: disease } }),
+                    signal: AbortSignal.timeout(10000),
+                });
+                if (!searchRes.ok)
+                    return `Open Targets error: HTTP ${searchRes.status}`;
+                const searchData = await searchRes.json();
+                const hits = searchData.data?.search?.hits || [];
+                const diseaseHits = hits.filter((h) => h.entity === 'disease');
+                if (diseaseHits.length === 0)
+                    return `No disease found for "${disease}" in Open Targets.`;
+                const output = [`## Disease Info: "${disease}"\n`];
+                for (const hit of diseaseHits.slice(0, 2)) {
+                    const diseaseId = hit.id;
+                    // Step 2: Get disease details with associations
+                    const detailQuery = `
+            query DiseaseDetail($id: String!) {
+              disease(efoId: $id) {
+                id
+                name
+                description
+                therapeuticAreas {
+                  id
+                  name
+                }
+                synonyms {
+                  terms
+                  relation
+                }
+                knownDrugs(size: 10) {
+                  uniqueTargetCount
+                  uniqueDrugCount
+                  rows {
+                    drug {
+                      id
+                      name
+                      drugType
+                      maximumClinicalTrialPhase
+                    }
+                    mechanismOfAction
+                    approvedIndications
+                  }
+                }
+                associatedTargets(page: { index: 0, size: 10 }) {
+                  count
+                  rows {
+                    target {
+                      id
+                      approvedSymbol
+                      approvedName
+                    }
+                    score
+                    datatypeScores {
+                      componentId: id
+                      score
+                    }
+                  }
+                }
+              }
+            }
+          `;
+                    const detailRes = await fetch('https://api.platform.opentargets.org/api/v4/graphql', {
+                        method: 'POST',
+                        headers: { 'Content-Type': 'application/json', 'User-Agent': UA },
+                        body: JSON.stringify({ query: detailQuery, variables: { id: diseaseId } }),
+                        signal: AbortSignal.timeout(10000),
+                    });
+                    if (!detailRes.ok)
+                        continue;
+                    const detailData = await detailRes.json();
+                    const d = detailData.data?.disease;
+                    if (!d)
+                        continue;
+                    let block = `### ${d.name}\n`;
+                    block += `**EFO ID:** [${d.id}](https://platform.opentargets.org/disease/${d.id})\n`;
+                    if (d.description) {
+                        const truncated = d.description.length > 500 ? d.description.slice(0, 500) + '...' : d.description;
+                        block += `\n${truncated}\n`;
+                    }
+                    // Therapeutic areas
+                    if (d.therapeuticAreas?.length > 0) {
+                        block += `\n**Therapeutic Areas:** ${d.therapeuticAreas.map((t) => t.name).join(', ')}\n`;
+                    }
+                    // Synonyms
+                    if (d.synonyms?.length > 0) {
+                        const exactSynonyms = d.synonyms
+                            .filter((s) => s.relation === 'HAS_EXACT_SYNONYM')
+                            .flatMap((s) => s.terms || [])
+                            .slice(0, 8);
+                        if (exactSynonyms.length > 0) {
+                            block += `**Synonyms:** ${exactSynonyms.join(', ')}\n`;
+                        }
+                    }
+                    // Top associated genes
+                    const targets = d.associatedTargets;
+                    if (targets?.rows?.length > 0) {
+                        block += `\n**Top Associated Genes** (${targets.count} total):\n`;
+                        for (const row of targets.rows.slice(0, 8)) {
+                            const t = row.target;
+                            block += `- **${t.approvedSymbol}** (${t.approvedName}) — score: ${row.score.toFixed(3)}\n`;
+                        }
+                    }
+                    // Known drugs
+                    const drugs = d.knownDrugs;
+                    if (drugs?.rows?.length > 0) {
+                        block += `\n**Known Drugs** (${drugs.uniqueDrugCount} drugs, ${drugs.uniqueTargetCount} targets):\n`;
+                        const seen = new Set();
+                        for (const row of drugs.rows) {
+                            const drugName = row.drug?.name || 'Unknown';
+                            if (seen.has(drugName))
+                                continue;
+                            seen.add(drugName);
+                            const phase = row.drug?.maximumClinicalTrialPhase ?? 'N/A';
+                            const moa = row.mechanismOfAction || '';
+                            block += `- **${drugName}** (phase ${phase}) — ${moa}\n`;
+                        }
+                    }
+                    output.push(block);
+                }
+                return output.join('\n---\n');
+            }
+            catch (e) {
+                return `Disease info lookup failed: ${e?.message || e}`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 11. Sequence Tools (Local analysis)
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'sequence_tools',
+        description: 'Local sequence analysis tools: GC content, reverse complement, translation (standard genetic code), ORF finding, motif search (regex), and molecular weight calculation. No API calls — runs instantly.',
+        parameters: {
+            sequence: { type: 'string', description: 'DNA/RNA/protein sequence (raw or FASTA format)', required: true },
+            operation: { type: 'string', description: 'Operation: "gc_content", "reverse_complement", "translate", "find_orfs", "motif_search", "molecular_weight"', required: true },
+            pattern: { type: 'string', description: 'Regex pattern for motif_search (e.g., "ATG[ATCG]{3,9}TAA")' },
+        },
+        tier: 'free',
+        async execute(args) {
+            const rawSeq = String(args.sequence).trim();
+            const operation = String(args.operation);
+            const pattern = args.pattern ? String(args.pattern) : '';
+            // Clean sequence: remove FASTA header, whitespace, numbers
+            const seq = rawSeq
+                .split('\n')
+                .filter(line => !line.startsWith('>'))
+                .join('')
+                .replace(/[\s\d]/g, '')
+                .toUpperCase();
+            if (seq.length === 0)
+                return 'No valid sequence provided. Remove FASTA headers and whitespace.';
+            const isDna = /^[ATCGN]+$/i.test(seq);
+            const isRna = /^[AUCGN]+$/i.test(seq);
+            const isProtein = /^[ACDEFGHIKLMNPQRSTVWY*]+$/i.test(seq) && !isDna;
+            switch (operation) {
+                case 'gc_content': {
+                    if (!isDna && !isRna)
+                        return 'GC content requires a DNA or RNA sequence (A, T/U, G, C, N).';
+                    const gc = (seq.match(/[GC]/gi) || []).length;
+                    const total = seq.replace(/N/gi, '').length;
+                    const pct = total > 0 ? ((gc / total) * 100).toFixed(2) : '0';
+                    const at = total - gc;
+                    let result = `## GC Content Analysis\n\n`;
+                    result += `**Sequence length:** ${seq.length} bp\n`;
+                    result += `**GC count:** ${gc} | **AT count:** ${at}\n`;
+                    result += `**GC%:** ${pct}%\n`;
+                    result += `**AT%:** ${(100 - parseFloat(pct)).toFixed(2)}%\n`;
+                    // Base composition
+                    const counts = {};
+                    for (const base of seq) {
+                        counts[base] = (counts[base] || 0) + 1;
+                    }
+                    result += `\n**Base composition:**\n`;
+                    for (const [base, count] of Object.entries(counts).sort()) {
+                        result += `- ${base}: ${count} (${((count / seq.length) * 100).toFixed(1)}%)\n`;
+                    }
+                    // Tm estimation (basic: 2*(A+T) + 4*(G+C) for short, or 64.9 + 41*(G+C-16.4)/N)
+                    if (seq.length <= 30) {
+                        const tm = 2 * at + 4 * gc;
+                        result += `\n**Estimated Tm (basic):** ${tm} C (for primers < 30 bp)\n`;
+                    }
+                    else {
+                        const tm = 64.9 + 41 * (gc - 16.4) / total;
+                        result += `\n**Estimated Tm (salt-adjusted):** ${tm.toFixed(1)} C\n`;
+                    }
+                    return result;
+                }
+                case 'reverse_complement': {
+                    if (!isDna && !isRna)
+                        return 'Reverse complement requires a DNA or RNA sequence.';
+                    const complementMap = isDna
+                        ? { A: 'T', T: 'A', G: 'C', C: 'G', N: 'N' }
+                        : { A: 'U', U: 'A', G: 'C', C: 'G', N: 'N' };
+                    const complement = seq.split('').map(b => complementMap[b] || 'N').join('');
+                    const revComp = complement.split('').reverse().join('');
+                    let result = `## Reverse Complement\n\n`;
+                    result += `**Input (${isDna ? 'DNA' : 'RNA'}, ${seq.length} bp):**\n`;
+                    result += `5'-\`${seq.length > 80 ? seq.slice(0, 40) + '...' + seq.slice(-40) : seq}\`-3'\n\n`;
+                    result += `**Complement:**\n`;
+                    result += `3'-\`${complement.length > 80 ? complement.slice(0, 40) + '...' + complement.slice(-40) : complement}\`-5'\n\n`;
+                    result += `**Reverse complement:**\n`;
+                    result += `5'-\`${revComp.length > 80 ? revComp.slice(0, 40) + '...' + revComp.slice(-40) : revComp}\`-3'\n`;
+                    return result;
+                }
+                case 'translate': {
+                    let dnaSeq = seq;
+                    if (isRna)
+                        dnaSeq = seq.replace(/U/g, 'T');
+                    if (!(/^[ATCGN]+$/i.test(dnaSeq)))
+                        return 'Translation requires a DNA or RNA sequence.';
+                    // Translate all 3 reading frames
+                    const results = [`## Translation (Standard Genetic Code)\n`];
+                    results.push(`**Input:** ${dnaSeq.length} bp\n`);
+                    for (let frame = 0; frame < 3; frame++) {
+                        const protein = [];
+                        for (let i = frame; i + 2 < dnaSeq.length; i += 3) {
+                            const codon = dnaSeq.slice(i, i + 3);
+                            if (codon.includes('N')) {
+                                protein.push('X');
+                            }
+                            else {
+                                protein.push(CODON_TABLE[codon] || 'X');
+                            }
+                        }
+                        const proteinStr = protein.join('');
+                        results.push(`**Frame +${frame + 1}:**`);
+                        results.push(`\`${proteinStr.length > 120 ? proteinStr.slice(0, 60) + '...' + proteinStr.slice(-60) : proteinStr}\``);
+                        results.push(`(${proteinStr.length} aa, ${(proteinStr.match(/\*/g) || []).length} stop codons)\n`);
+                    }
+                    // Highlight first ORF in frame +1
+                    const frame1 = [];
+                    for (let i = 0; i + 2 < dnaSeq.length; i += 3) {
+                        const codon = dnaSeq.slice(i, i + 3);
+                        frame1.push(codon.includes('N') ? 'X' : (CODON_TABLE[codon] || 'X'));
+                    }
+                    const protStr = frame1.join('');
+                    const orfMatch = protStr.match(/M[^*]+/);
+                    if (orfMatch) {
+                        results.push(`**Longest ORF (frame +1):** ${orfMatch[0].length} aa starting at M`);
+                        results.push(`\`${orfMatch[0].slice(0, 80)}${orfMatch[0].length > 80 ? '...' : ''}\``);
+                    }
+                    return results.join('\n');
+                }
+                case 'find_orfs': {
+                    let dnaSeq = seq;
+                    if (isRna)
+                        dnaSeq = seq.replace(/U/g, 'T');
+                    if (!(/^[ATCGN]+$/i.test(dnaSeq)))
+                        return 'ORF finding requires a DNA or RNA sequence.';
+                    const minOrfLength = 30; // minimum 30 aa = 90 bp
+                    const orfs = [];
+                    // Search all 3 forward frames
+                    for (let frame = 0; frame < 3; frame++) {
+                        const protein = [];
+                        for (let i = frame; i + 2 < dnaSeq.length; i += 3) {
+                            const codon = dnaSeq.slice(i, i + 3);
+                            protein.push(codon.includes('N') ? 'X' : (CODON_TABLE[codon] || 'X'));
+                        }
+                        const protStr = protein.join('');
+                        // Find all M...* ORFs
+                        const orfRe = /M[^*]*/g;
+                        let m;
+                        while ((m = orfRe.exec(protStr)) !== null) {
+                            if (m[0].length >= minOrfLength) {
+                                const aaStart = m.index;
+                                orfs.push({
+                                    frame: frame + 1,
+                                    start: frame + aaStart * 3 + 1, // 1-based nucleotide position
+                                    end: frame + (aaStart + m[0].length) * 3,
+                                    length: m[0].length,
+                                    protein: m[0],
+                                });
+                            }
+                        }
+                    }
+                    // Also search reverse complement
+                    const compMap = { A: 'T', T: 'A', G: 'C', C: 'G', N: 'N' };
+                    const rcSeq = dnaSeq.split('').map(b => compMap[b] || 'N').reverse().join('');
+                    for (let frame = 0; frame < 3; frame++) {
+                        const protein = [];
+                        for (let i = frame; i + 2 < rcSeq.length; i += 3) {
+                            const codon = rcSeq.slice(i, i + 3);
+                            protein.push(codon.includes('N') ? 'X' : (CODON_TABLE[codon] || 'X'));
+                        }
+                        const protStr = protein.join('');
+                        const orfRe = /M[^*]*/g;
+                        let m;
+                        while ((m = orfRe.exec(protStr)) !== null) {
+                            if (m[0].length >= minOrfLength) {
+                                const aaStart = m.index;
+                                orfs.push({
+                                    frame: -(frame + 1),
+                                    start: dnaSeq.length - (frame + (aaStart + m[0].length) * 3) + 1,
+                                    end: dnaSeq.length - (frame + aaStart * 3),
+                                    length: m[0].length,
+                                    protein: m[0],
+                                });
+                            }
+                        }
+                    }
+                    // Sort by length descending
+                    orfs.sort((a, b) => b.length - a.length);
+                    let result = `## Open Reading Frames (min 30 aa)\n\n`;
+                    result += `**Sequence length:** ${dnaSeq.length} bp\n`;
+                    result += `**ORFs found:** ${orfs.length}\n\n`;
+                    if (orfs.length === 0) {
+                        result += 'No ORFs >= 30 aa found. Try a longer sequence or lower threshold.\n';
+                    }
+                    else {
+                        for (const orf of orfs.slice(0, 15)) {
+                            result += `**Frame ${orf.frame > 0 ? '+' : ''}${orf.frame}** | nt ${orf.start}-${orf.end} | **${orf.length} aa** (${orf.length * 3} bp)\n`;
+                            result += `\`${orf.protein.slice(0, 60)}${orf.protein.length > 60 ? '...' : ''}\`\n\n`;
+                        }
+                    }
+                    return result;
+                }
+                case 'motif_search': {
+                    if (!pattern)
+                        return 'Motif search requires a pattern parameter (regex). Example: "ATG[ATCG]{3,9}TAA"';
+                    let re;
+                    try {
+                        re = new RegExp(pattern, 'gi');
+                    }
+                    catch (err) {
+                        return `Invalid regex pattern: ${err?.message || err}`;
+                    }
+                    const matches = [];
+                    let m;
+                    while ((m = re.exec(seq)) !== null) {
+                        matches.push({ start: m.index + 1, end: m.index + m[0].length, match: m[0] });
+                        // Prevent infinite loop on zero-length matches
+                        if (m[0].length === 0)
+                            re.lastIndex++;
+                    }
+                    let result = `## Motif Search\n\n`;
+                    result += `**Pattern:** \`${pattern}\`\n`;
+                    result += `**Sequence length:** ${seq.length}\n`;
+                    result += `**Matches found:** ${matches.length}\n\n`;
+                    if (matches.length === 0) {
+                        result += 'No matches found.\n';
+                    }
+                    else {
+                        for (const match of matches.slice(0, 50)) {
+                            const display = match.match.length > 60 ? match.match.slice(0, 60) + '...' : match.match;
+                            result += `- **Position ${match.start}-${match.end}:** \`${display}\`\n`;
+                        }
+                        if (matches.length > 50)
+                            result += `\n... and ${matches.length - 50} more matches.\n`;
+                    }
+                    return result;
+                }
+                case 'molecular_weight': {
+                    let result = `## Molecular Weight\n\n`;
+                    result += `**Sequence length:** ${seq.length}\n`;
+                    if (isProtein) {
+                        // Protein MW: sum of AA weights - (n-1) * water (18.02)
+                        let mw = 0;
+                        let unknowns = 0;
+                        for (const aa of seq) {
+                            if (aa === '*')
+                                continue; // stop codon
+                            if (AA_WEIGHTS[aa]) {
+                                mw += AA_WEIGHTS[aa];
+                            }
+                            else {
+                                unknowns++;
+                                mw += 128.16; // average AA weight
+                            }
+                        }
+                        // Subtract water for peptide bonds
+                        const aas = seq.replace(/\*/g, '').length;
+                        mw -= (aas - 1) * 18.02;
+                        result += `**Type:** Protein (${aas} amino acids)\n`;
+                        result += `**Molecular Weight:** ${mw.toFixed(2)} Da (${(mw / 1000).toFixed(2)} kDa)\n`;
+                        if (unknowns > 0)
+                            result += `*Note: ${unknowns} unknown residues estimated at 128.16 Da (average)*\n`;
+                        // Extinction coefficient estimate (Pace method)
+                        const nTrp = (seq.match(/W/g) || []).length;
+                        const nTyr = (seq.match(/Y/g) || []).length;
+                        const nCys = (seq.match(/C/g) || []).length;
+                        const e280 = nTrp * 5500 + nTyr * 1490 + nCys * 125;
+                        result += `\n**Extinction coefficient (280nm):** ${e280} M\u207B\u00B9cm\u207B\u00B9\n`;
+                        result += `(${nTrp} Trp, ${nTyr} Tyr, ${nCys} Cys)\n`;
+                        // Isoelectric point estimate (very rough)
+                        const nAsp = (seq.match(/D/g) || []).length;
+                        const nGlu = (seq.match(/E/g) || []).length;
+                        const nHis = (seq.match(/H/g) || []).length;
+                        const nLys = (seq.match(/K/g) || []).length;
+                        const nArg = (seq.match(/R/g) || []).length;
+                        const negCharge = nAsp + nGlu;
+                        const posCharge = nHis + nLys + nArg;
+                        result += `\n**Charge residues:** ${posCharge} positive (K:${nLys} R:${nArg} H:${nHis}), ${negCharge} negative (D:${nAsp} E:${nGlu})\n`;
+                    }
+                    else if (isDna || isRna) {
+                        // Nucleic acid MW
+                        const weights = isDna ? DNA_WEIGHTS : { A: 347.2, U: 324.2, G: 363.2, C: 323.2 };
+                        let mw = 0;
+                        for (const base of seq) {
+                            mw += weights[base] || 330; // average for N
+                        }
+                        // Subtract water for phosphodiester bonds, add 5' phosphate
+                        mw -= (seq.length - 1) * 18.02;
+                        result += `**Type:** ${isDna ? 'DNA' : 'RNA'} (${seq.length} ${isDna ? 'bp' : 'nt'})\n`;
+                        result += `**Molecular Weight (ss):** ${mw.toFixed(2)} Da (${(mw / 1000).toFixed(2)} kDa)\n`;
+                        if (isDna) {
+                            result += `**Molecular Weight (ds):** ~${(mw * 2).toFixed(0)} Da (${((mw * 2) / 1000).toFixed(2)} kDa)\n`;
+                        }
+                        // Concentration conversion
+                        const ugPerOd = isDna ? (seq.length < 25 ? 33 : 50) : 40;
+                        result += `\n**1 OD260 =** ~${ugPerOd} ug/mL\n`;
+                    }
+                    else {
+                        result += 'Could not determine sequence type (DNA/RNA/protein). Check for invalid characters.\n';
+                    }
+                    return result;
+                }
+                default:
+                    return `Unknown operation: "${operation}". Supported: gc_content, reverse_complement, translate, find_orfs, motif_search, molecular_weight`;
+            }
+        },
+    });
+    // ════════════════════════════════════════════════════════════════════════
+    // 12. Ecology Data (GBIF Occurrences)
+    // ════════════════════════════════════════════════════════════════════════
+    registerTool({
+        name: 'ecology_data',
+        description: 'Search GBIF for biodiversity occurrence records. Find species observations with location, date, and collection data.',
+        parameters: {
+            species: { type: 'string', description: 'Scientific name (e.g., "Panthera tigris", "Quercus robur")', required: true },
+            country: { type: 'string', description: 'ISO 3166-1 alpha-2 country code (e.g., "US", "BR", "AU")' },
+            limit: { type: 'number', description: 'Max records (default 20, max 100)' },
+        },
+        tier: 'free',
+        async execute(args) {
+            const species = String(args.species);
+            const country = args.country ? String(args.country).toUpperCase() : '';
+            const limit = Math.min(typeof args.limit === 'number' ? args.limit : 20, 100);
+            try {
+                let url = `https://api.gbif.org/v1/occurrence/search?scientificName=${encodeURIComponent(species)}&limit=${limit}&hasCoordinate=true`;
+                if (country)
+                    url += `&country=${country}`;
+                const res = await fetch(url, {
+                    headers: { 'User-Agent': UA },
+                    signal: AbortSignal.timeout(10000),
+                });
+                if (!res.ok)
+                    return `GBIF error: HTTP ${res.status}`;
+                const data = await res.json();
+                const results_arr = data.results || [];
+                const totalCount = data.count || 0;
+                if (results_arr.length === 0)
+                    return `No GBIF occurrence records for "${species}"${country ? ` in ${country}` : ''}. Try the full scientific name.`;
+                const output = [`## GBIF Occurrences: *${species}*\n`];
+                output.push(`**Total records:** ${totalCount.toLocaleString()}${country ? ` (filtered: ${country})` : ''}\n`);
+                // Summary stats
+                const countries = new Map();
+                const years = new Map();
+                const basisOfRecord = new Map();
+                for (const rec of results_arr) {
+                    const c = rec.country || 'Unknown';
+                    countries.set(c, (countries.get(c) || 0) + 1);
+                    if (rec.year)
+                        years.set(rec.year, (years.get(rec.year) || 0) + 1);
+                    const basis = rec.basisOfRecord || 'Unknown';
+                    basisOfRecord.set(basis, (basisOfRecord.get(basis) || 0) + 1);
+                }
+                // Country distribution
+                const sortedCountries = [...countries.entries()].sort((a, b) => b[1] - a[1]);
+                output.push(`**Countries in sample:** ${sortedCountries.map(([c, n]) => `${c} (${n})`).join(', ')}\n`);
+                // Record types
+                output.push(`**Record types:** ${[...basisOfRecord.entries()].map(([t, n]) => `${t.replace(/_/g, ' ')} (${n})`).join(', ')}\n`);
+                // Year range
+                const yearKeys = [...years.keys()].sort();
+                if (yearKeys.length > 0) {
+                    output.push(`**Year range:** ${yearKeys[0]}–${yearKeys[yearKeys.length - 1]}\n`);
+                }
+                // Individual records
+                output.push('\n### Records\n');
+                for (const rec of results_arr.slice(0, 20)) {
+                    const name = rec.species || rec.scientificName || species;
+                    const lat = rec.decimalLatitude?.toFixed(4) || '?';
+                    const lon = rec.decimalLongitude?.toFixed(4) || '?';
+                    const date = rec.eventDate || rec.year || 'N/A';
+                    const loc = rec.locality || rec.stateProvince || '';
+                    const countryName = rec.country || '';
+                    const institution = rec.institutionCode || '';
+                    const basis = rec.basisOfRecord?.replace(/_/g, ' ') || '';
+                    let line = `- **${name}** — ${lat}, ${lon}`;
+                    if (countryName)
+                        line += ` (${countryName}${loc ? `, ${loc}` : ''})`;
+                    if (date)
+                        line += ` | ${date}`;
+                    if (basis)
+                        line += ` | ${basis}`;
+                    if (institution)
+                        line += ` | ${institution}`;
+                    if (rec.gbifID)
+                        line += ` | [GBIF:${rec.gbifID}](https://www.gbif.org/occurrence/${rec.gbifID})`;
+                    output.push(line);
+                }
+                return output.join('\n');
+            }
+            catch (e) {
+                return `Ecology data lookup failed: ${e?.message || e}`;
+            }
+        },
+    });
+}
+//# sourceMappingURL=lab-bio.js.map