skills.rest
K-Dense-AI
K-Dense-AIOFFICIAL
Explore all skills by this author →
Website𝕏@k_dense_aiGitHub

genomic-intelligence

Predict promoters, splice sites, enhancers, chromatin state, and gene expression from DNA sequences via hosted transformer models.

41.1k|3.8k|Updated Oct 19, 2025
Search Tags:#genomics#REST api#DNA sequence#MCP Server#promoter prediction#gene expression prediction#splice sites
One-click install
npx skills add https://github.com/K-Dense-AI/scientific-agent-skills --skill genomic-intelligence
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: genomic-intelligence
Source: https://github.com/K-Dense-AI/scientific-agent-skills/tree/main/skills/genomic-intelligence
Command: npx skills add https://github.com/K-Dense-AI/scientific-agent-skills --skill genomic-intelligence

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires requests, and includes references (resource) components.

What problem does it solve?

Running DNA language model inference normally requires local GPUs, model weights, and heavy Python stacks. This Skill lets you predict regulatory features, gene structure, and expression directly from a DNA sequence, gene symbol, or genomic region through Genomic Intelligence's hosted REST API and keyless MCP server, with nothing running locally.

Core Features & Use Cases

  • Six sequence-analysis tasks: promoter region detection, splice donor/acceptor prediction, enhancer activity scoring (DeepSTARR), chromatin state annotation (DeepSEA), sequence-to-expression prediction (log TPM), and de-novo gene annotation.
  • Two access paths: a keyless hosted MCP server with handle-based tools that keep large sequences out of context, and a universal REST API authenticated with a GI_API_KEY bearer token.
  • Composite workflow: find genes in a region and predict each gene's expression in one call, with Ensembl-based sequence acquisition helpers that build the exact 9,198 bp TSS-centred window the expression model requires.
  • Use Case: Given the gene HBB, fetch its TSS-centred window and predict its expression in K562 cells as log(TPM+1), or submit a 20 kb region for async de-novo transcript annotation and poll for the result.

Quick Start

Ask the agent to use the Genomic Intelligence MCP server to load the demo sequence promoter_tp53 and predict promoter regions in it.

Frequently Asked Questions about genomic-intelligence

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

FAQPage Schema
How do I predict gene expression from a DNA sequence?▼

Use the expression task with exactly 9,198 bp centred on the TSS plus a cell-type description such as "K562 cells". On MCP, call fetch_gene_for_expression to build the window, then predict_expression; over REST, POST to /v1/tasks/expression/predict with options.description set.

How do I find genes in a genomic region without reference annotation?▼

Use the annotation task, which predicts de-novo transcripts from sequence alone. On MCP, acquire a handle with fetch_region and call find_genes; over REST, submit to /v1/tasks/annotation/predict with Prefer: respond-async and poll the returned job_id until it completes.

Do I need an API key to use Genomic Intelligence?▼

The hosted MCP server at mcp.genomicintelligence.ai/mcp works keyless against a capped public demo quota, so no key is needed to try it. The REST /v1 API requires a GI_API_KEY bearer token, which you can request from [email protected].

Why does the expression prediction return a 422 validation error?▼

The expression model rejects any sequence that is not exactly 9,198 bp centred on the TSS, or any request missing the required options.description cell-type string. Build the window with fetch_gene_for_expression or the documented TSS-centring math rather than truncating by hand.

Can I use Genomic Intelligence for non-human species?▼

Yes, non-human models exist for several tasks including Drosophila, yeast, and Arabidopsis, and the enhancer default (DeepSTARR) is a Drosophila model. Discover available model IDs at call time with GET /v1/tasks/{task}/models or list_models, and use Ensembl production names like mus_musculus for sequence acquisition.

When should I not use hosted DNA sequence inference?▼

Do not use it for local alignment, variant calling, or file I/O, which belong to local tools like BioPython or bcftools. It is intended for model inference from sequence for research and development, not clinical or diagnostic decisions.

Related Skills

Explore More Education & Research Skills→
SciMate-AISciMate-AI
1

ensembl-database

Query the Ensembl REST API for gene, sequence, variant, and regulatory data.

Official
Advanced
NeuroAIHubNeuroAIHub
269

gnomad-database

Query gnomAD for allele frequencies, gene constraint metrics, and variant lists.

Official
Intermediate
gabrielvuksanigabrielvuksani

bgpt-paper-search

Search a curated scientific database for structured experimental data from full-text papers.

Community
Intermediate
RepairYourTechRepairYourTech
6

ollama

Run open-source LLMs locally via the Ollama runtime REST API.

Community
Intermediate