Skip to main content
Glama

search_similar_structures

Identify structurally similar chemicals in SureChEMBL by inputting a reference chemical ID and adjusting similarity thresholds. Retrieve relevant results for research or analysis.

Instructions

Find structurally similar chemicals using similarity search

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
limitNoNumber of results to return (1-100, default: 25)
reference_idYesReference chemical ID for similarity search
thresholdNoSimilarity threshold (0.0-1.0, default: 0.7)

Implementation Reference

  • Handler function implementing the tool logic: validates args, fetches reference chemical, constructs a mock similarity search response explaining API limitations and providing alternatives.
    private async handleSearchSimilarStructures(args: any) { if (!args || typeof args.reference_id !== 'string') { throw new McpError(ErrorCode.InvalidParams, 'Invalid reference chemical ID'); } try { // Get the reference chemical first const refResponse = await this.apiClient.get(`/chemical/id/${args.reference_id}`); const refChemical = refResponse.data.data?.[0]; if (!refChemical) { throw new Error('Reference chemical not found'); } // Since SureChEMBL doesn't have direct similarity search, we'll provide a mock implementation // that explains the limitation and suggests alternatives const similarityResult = { reference_chemical: { id: args.reference_id, name: refChemical.name, smiles: refChemical.smiles, molecular_weight: refChemical.mol_weight }, search_parameters: { threshold: args.threshold || 0.7, limit: args.limit || 25 }, message: 'Direct similarity search not available in SureChEMBL API', suggestions: [ 'Use chemical name variations to find related compounds', 'Search by molecular weight ranges', 'Use external cheminformatics tools for similarity search', 'Try searching by chemical class or functional groups' ], alternative_searches: { by_name_fragments: `Try searching for fragments of "${refChemical.name}"`, by_molecular_weight: `Search for compounds with molecular weight around ${refChemical.mol_weight}`, by_chemical_class: 'Search for compounds in the same chemical class' } }; return { content: [ { type: 'text', text: JSON.stringify(similarityResult, null, 2), }, ], }; } catch (error) { throw new McpError( ErrorCode.InternalError, `Failed to search similar structures: ${error instanceof Error ? error.message : 'Unknown error'}` ); } }
  • JSON schema defining input parameters for the tool: reference_id (required), threshold, limit.
    inputSchema: { type: 'object', properties: { reference_id: { type: 'string', description: 'Reference chemical ID for similarity search' }, threshold: { type: 'number', description: 'Similarity threshold (0.0-1.0, default: 0.7)', minimum: 0.0, maximum: 1.0 }, limit: { type: 'number', description: 'Number of results to return (1-100, default: 25)', minimum: 1, maximum: 100 }, }, required: ['reference_id'], },
  • src/index.ts:510-522 (registration)
    Tool registration in the setTools array, including name, description, and schema.
    { name: 'search_similar_structures', description: 'Find structurally similar chemicals using similarity search', inputSchema: { type: 'object', properties: { reference_id: { type: 'string', description: 'Reference chemical ID for similarity search' }, threshold: { type: 'number', description: 'Similarity threshold (0.0-1.0, default: 0.7)', minimum: 0.0, maximum: 1.0 }, limit: { type: 'number', description: 'Number of results to return (1-100, default: 25)', minimum: 1, maximum: 100 }, }, required: ['reference_id'], }, },
  • src/index.ts:574-575 (registration)
    Dispatch case in the CallToolRequestSchema handler that routes to the tool handler.
    case 'search_similar_structures': return await this.handleSearchSimilarStructures(args);

Other Tools

Related Tools

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Augmented-Nature/SureChEMBL-MCP-Server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server