ngs-analysis Codex Plugin
Plugin Codex Education & Research Data Engineering & AnalyticsAI Agents & AI App DevelopmentClaude Code Customization & WorkflowThe explanation below is AI-generated. Please verify it against the sources.
ngs-analysis is a Codex plugin (version 1.0.3) by OpenAI that provides guided intake, routing, and local execution for next-generation sequencing (NGS) workflows, covering BCL, FASTQ, DNA variant calling, RNA-seq, single-cell, epigenomics, amplicon, and metagenomics analyses. According to the plugin.json and README, it inspects local sequencing inputs first, asks only the minimum assay-specific questions, checks whether required tools already exist before suggesting any install, and prefers public or freely accessible tools and nf-core workflows where practical. It ships 18 named skills (e.g. ngs-analysis-router, ngs-fastq-qc, ngs-dna-germline-variants, ngs-bulk-rnaseq-counts-qc, ngs-scrna-seq, ngs-shotgun-metagenomics) and a set of local execution scripts that write standardized run manifests, validation reports, logs, QC summaries, and checksummed artifact indexes. The README states capability status is 'intentionally mixed maturity': some lanes (FASTQ QC, bulk RNA-seq counts/QC and differential expression, scRNA-seq FASTQ-to-count and post-count QC, generic/germline DNA variants, epigenomics/amplicon/shotgun FASTQ QC, BCL-to-FASTQ) run locally, while others act as dispatchers that hand off to public pipelines such as nf-core/sarek for full workflows. It also includes an nf-core adapter and reference/database readiness tooling (ngs_reference_manager.py) to check or plan required genome/database bundles before execution.
Overview
ngs-analysis is not a hosted SaaS; per the source it is a Codex plugin package distributed as a marketplace root (.agents/plugins/marketplace.json plus plugins/ngs-analysis/) that a user installs locally as a Codex marketplace entry, after which it is invoked by name or by referencing its included skills.
What you can do with ngs-analysis
- Get guided intake that inspects local BCL/FASTQ files or count matrices and asks only the missing assay-specific questions before routing to a workflow
- Run local execution lanes for FASTQ QC (FastQC/MultiQC/trimming), bulk RNA-seq counts/QC (Salmon) and differential expression (DESeq2/edgeR/limma), scRNA-seq FASTQ-to-count (STARsolo) and post-count QC (scanpy-based), generic/germline DNA variant calling (samtools/bcftools/GATK), and BCL-to-FASTQ conversion (bcl-convert/bcl2fastq)
- Use FASTQ validation/QC packages for epigenomics (ATAC-seq, ChIP-seq, CUT&RUN, CUT&Tag), amplicon microbiome (16S/18S/ITS/COI via QIIME2 or DADA2), and shotgun metagenomics (Kraken2/Bracken/HUMAnN)
- Check tool availability with a preflight script (ngs_preflight.py) that inspects PATH, optionally queries package indexes/registries, and generates a reviewable install plan without installing anything unless explicitly approved
- Check or plan reference/database readiness with ngs_reference_manager.py (list, check, plan, setup-plan, inventory, lock, verify-lock)
- Hand off full workflows to public nf-core pipelines (rnaseq, scrnaseq, sarek, atacseq, chipseq, cutandrun, ampliseq, taxprofiler) via a dedicated adapter that generates pinned Nextflow commands
- Receive standardized run outputs for each execution: run manifests, validation summaries, logs, QC reports, checksummed artifact indexes, and visualization bundles
Sources
Original description (English)
Guided NGS intake, local execution, and public-pipeline routing for BCL, FASTQ, DNA variant, RNA-seq, single-cell, epigenomics, amplicon, and metagenomics analyses, with deeper decision skills for high-risk assay branches.
History of ngs-analysis
- Codex Plugin Added ngs-analysis