protrek-search

Performs tri-modal searches across sequence, structure, and text using ProTrek API and gradio_client.

1|Updated Apr 29, 2026
One-click install
npx skills add https://github.com/cultivarium/hermes-starter-kit --skill protrek-search
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: protrek-search
Source: https://github.com/cultivarium/hermes-starter-kit/tree/main/skills/protrek-search
Command: npx skills add https://github.com/cultivarium/hermes-starter-kit --skill protrek-search

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Retrieve semantically relevant protein entries by querying across sequence, structure, and free-text descriptions, enabling cross-modal discovery without relying on exact sequence matches.

Core Features & Use Cases

  • Tri-modal search across sequence, structure, and text via a hosted ProTrek API.
  • Supports sequence → sequence, sequence → text, and text → sequence/structure queries for annotation, homolog discovery, and cross-modal retrieval across Swiss-Prot and related databases.
  • Use case: discover proteins with similar function to a description or annotate an unknown sequence with functional terms.

Quick Start

Query ProTrek with a sequence or description to retrieve similar proteins and functional annotations.

Frequently Asked Questions about protrek-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I find proteins with similar function using a text description instead of a sequence?

You can find proteins with similar function from a text description by using cross-modal semantic search. This Skill maps text queries directly to sequence and structure databases, retrieving functionally related protein entries without requiring exact sequence matches.

Can I annotate an unknown protein sequence with functional terms automatically?

Yes, you can annotate an unknown protein sequence with functional terms by running a sequence-to-text cross-modal query. It searches across Swiss-Prot and related databases to retrieve semantic functional annotations for your input sequence.

What is cross-modal protein search and how does it differ from BLAST sequence alignment?

Cross-modal protein search identifies functionally related proteins across sequence, structure, and text using semantic retrieval. Unlike exact sequence alignment, it enables querying across different data types, such as finding structural proteins from text descriptions.

Do I need to install the gradio_client to query the ProTrek database?

Yes, you need the gradio_client to call the hosted ProTrek API at search-protrek.com. The Skill uses this client to access the /search endpoint with your specified input type, output type, and database selections.

How do I discover homologous proteins across Swiss-Prot using semantic sequence search?

You can discover homologous proteins by running a sequence-to-sequence semantic search query. This Skill compares your input sequence against the Swiss-Prot database to retrieve functionally related and homologous entries without relying on exact string matches.