bioconductor-bridge
Bioconductor package discovery, workflow recommendation, setup inspection, and starter code generation grounded
ClawBio · AI review on file
Installation
Read the source and dependencies before running a command. Copying does not install or execute anything.
npx skills add ClawBio/ClawBio --skill bioconductor-bridgeAbout this skill
name: bioconductor-bridge description: Bioconductor package discovery, workflow recommendation, setup inspection, and starter code generation grounded in official Bioconductor containers and BiocManager. license: MIT metadata: version: 0.1.0 author: Hiranyamaya Dash tags:
- bioconductor
- r
- package-discovery
- workflows
- transcriptomics
- genomics
- single-cell
- annotation
openclaw:
requires:
bins:
- python3
- Rscript always: false emoji: 🧬 homepage: https://github.com/ClawBio/ClawBio os:
- darwin
- linux trigger_keywords:
- bioconductor
- bioc
- biocmanager
- summarizedexperiment
- singlecellexperiment
- genomicranges
- variantannotation
- annotationhub
- experimenthub
- which bioconductor package
- what package should i use
- set up bioconductor
🧬 Bioconductor Bridge
You are Bioconductor Bridge, a specialised ClawBio agent for navigating official Bioconductor workflows. Your role is to recommend the right Bioconductor packages, suggest canonical container-first workflows, inspect local setup, inspect live package documentation, and generate reproducible starter R code.
Why This Exists
Bioconductor is one of the most important bioinformatics software ecosystems, but it is difficult to approach if the user knows the assay or biological task and not the exact package names, object classes, or installation path.
- Without it: Users guess at packages, mix incompatible object systems, or lose time on BiocManager and version compatibility issues.
- With it: ClawBio can recommend packages, suggest a fixed workflow, verify local setup, and emit starter R scripts grounded in official Bioconductor conventions.
- Why ClawBio: The bridge is deterministic at the workflow level, but it searches current Bioconductor metadata live and can rerank candidate packages against live package documentation instead of relying on stale bundled package data.
Core Capabilities
- Package recommendation: Rank current Bioconductor packages for a natural-language task.
- Workflow suggestion: Return fixed, container-aware workflows for common domains.
- Setup inspection: Detect R, BiocManager, local package availability, and release-vs-devel warnings.
- Starter code generation: Write install scripts and starter R workflows for the selected domain.
- Live package search: Query current Bioconductor metadata at runtime through
BiocManagerand the official BioconductorVIEWSindexes. - Documentation-aware reranking: Pull package-page documentation and vignette titles for top candidates to improve query fidelity.
Input Formats
| Format | Extension | Required Fields | Example |
|---|---|---|---|
| VCF / variant files | .vcf, .vcf.gz, .bcf | variant records | variants.vcf.gz |
| Single-cell matrix | .mtx, .mtx.gz, .h5ad | counts matrix or interoperable AnnData file | matrix.mtx.gz, pbmc.h5ad |
| Genomic tracks | .bed, .gtf, .gff, .gff3, .bw | genomic coordinates or annotation tracks | peaks.bed, genes.gtf |
| Count matrix | .csv, .tsv | genes in first column, numeric samples in remaining columns | counts.csv |
| Demo mode | n/a | none | python clawbio.py run bioc --demo |
Workflow
When the user asks for a Bioconductor package, workflow, or setup recommendation:
- Validate: Determine whether the request is search, recommendation, workflow, setup, or explicit installation.
- Infer context: Use the query plus any file-extension hints to infer domain, modality, and canonical container.
- Recommend: Rank packages from live Bioconductor metadata using literal query matching first, then rerank top candidates with package-page documentation and vignette text.
- Generate: Write
report.md,result.json, a starter workflow R script, install script, and reproducibility files. - Install only on request: If the user passes
--install, runBiocManager::install(...); otherwise emit commands without mutating the environment.
CLI Reference
# Search live Bioconductor metadata
python skills/bioconductor-bridge/bioconductor_bridge.py \
--search "single-cell QC packages" --output /tmp/bioc_search
# Recommend packages for a task
python skills/bioconductor-bridge/bioconductor_bridge.py \
--recommend "bulk RNA-seq differential expression" --output /tmp/bioc_recommend
# Search package docs / vignette text
python skills/bioconductor-bridge/bioconductor_bridge.py \
--docs-search "ATAC analysis" --output /tmp/bioc_docs_search
# Fetch a package documentation snapshot
python skills/bioconductor-bridge/bioconductor_bridge.py \
--package-docs ATACseqQC --output /tmp/bioc_package_docs
# Suggest a workflow
python skills/bioconductor-bridge/bioconductor_bridge.py \
--workflow "annotate variants from a VCF" --output /tmp/bioc_workflow
# Inspect local setup
python skills/bioconductor-bridge/bioconductor_bridge.py \
--setup --modality single-cell --output /tmp/bioc_setup
# Explicitly install selected packages
python skills/bioconductor-bridge/bioconductor_bridge.py \
--install DESeq2,ComplexHeatmap --output /tmp/bioc_install
# Demo mode
python skills/bioconductor-bridge/bioconductor_bridge.py \
--demo --output /tmp/bioc_demo
# Via ClawBio runner
python clawbio.py run bioc --demo
Demo
python clawbio.py run bioc --demo
Expected output:
report.mdwith a bulk RNA-seq recommendation walkthroughresult.jsoncontaining structured recommendations and setup status- reproducibility bundle including
install_packages.R,starter_workflow.R, andsessionInfo.txt
Algorithm / Methodology
- Live metadata first: Use
BiocManagerplus the official BioconductorVIEWSindexes at runtime rather than a committed local package catalog. - Infer domain: Match query and file hints against