bioinfo-lab

Plugin bioinfo-lab from marketplace research_agents in oumarkme/research_agents. Commands read from the repo's manifest — today.

Skills (84)

• access-benefit-sharing-and-data-governance

  Use when you sample, sequence, or share data from wild, germplasm, microbial, or community-held material across borders. Not for consent and authorship (research-ethics) or repository choice (data-stewardship).

• agent-brief

  Use when you write the prompt for a team agent or subagent: the brief template, parallelism rules, and the checks to run on each return.

• agricultural-microbiome

  Use when planning or reporting rumen, soil, or rhizosphere microbiome studies: sampling, controls, amplicon versus shotgun, host reads, reporting standards. Not for read-level analysis (amplicon-metabarcoding-analysis, metagenomics).

• ai-tools-in-analysis

  Use when an AI coding or writing assistant helps with analysis code, numbers, or text: what to delegate, verification, provenance, disclosure. Not for editing prose tells (humanizer).

• amplicon-metabarcoding-analysis

  Use when analyzing 16S, 18S, ITS, or other amplicon or eDNA metabarcoding reads: ASVs, taxonomy, controls, contamination, reporting. Not for shotgun metagenomes (metagenomics) or sampling planning (sample-and-assay-planning).

• amr-genotype-phenotype

  Use when detecting antimicrobial resistance genes and mutations in bacterial genomes and relating them to phenotype. Not for building the assembly (bacterial-genome-assembly-annotation) or core-genome work (pangenome-analysis-prokaryote).

• analysis-log

  Use when work spans several steps or sessions (pipeline build, multi-step analysis, long jobs, manuscript revision): progress file, claims ledger, start-of-session routine, and provenance for every reported number.

• animal-research-ethics-3rs

  Use when planning or reporting animal experiments or field and wildlife sampling: 3Rs, reporting guideline, sample size, approvals, and permits. Not for human data (research-ethics) or general power analysis (study-design).

• annotation-and-databases

  Use when annotating variants in any organism: consequence tools, transcript sets, frequency, constraint, and clinical database releases, or species databases. Not for classification (variant-interpretation).

• array-genotyping-and-imputation

  Use when processing SNP-array or low-pass data in any organism: array QC, strand and build, phasing, imputation, panel choice, info filters. Not for GWAS models (association-studies) or sample swaps (genomics-pitfalls).

• association-studies

  Use when running or reviewing GWAS, QTL mapping, or Mendelian randomization in humans, crops, or livestock. Not for general design (study-design) or prediction models (genomic-prediction).

• bacterial-genome-assembly-annotation

  Use when assembling, polishing, annotating, and checking a bacterial isolate genome from long or hybrid reads. Not for AMR calls (amr-genotype-phenotype) or many-genome comparison (pangenome-analysis-prokaryote).

• breeding-metadata-and-archives

  Use when describing or sharing phenotype, sample, and sequence data from breeding or agricultural projects (MIAPPE, BrAPI, ontologies, FAANG, ENA, EVA). Not for general data layout or DMPs (data-stewardship).

• clinical-pipeline-validation

  Use when validating, freezing, or changing a human diagnostic NGS bioinformatics pipeline, or preparing evidence for accreditation. Not for variant classification (variant-interpretation) or research benchmarking (sequencing-pipelines).

• coding-style

  Use before writing, changing, or reviewing code in Python, R, C++, Rust, shell, LaTeX, Nextflow, or Snakemake: Google style, project layout, tests, reproducibility.

• current-methods-lookup

  Use when a decision depends on version-sensitive facts: choosing a tool, options, thresholds, database releases, benchmarks, or guidelines. Not for stable principles already covered by the domain skill.

• data-stewardship

  Use when planning, organizing, documenting, or sharing research data in any organism: layout, metadata, provenance, FAIR, repositories, data management plans. Not for analysis.

• deep-learning-genomics

  Use when training, validating, or applying neural models to genomic data (variant effect predictors, sequence models, basecallers). Not for polygenic or genomic prediction (genomic-prediction) or ACMG classification (variant-interpretation).

• difficult-genes-and-mito

  Use when a gene has a paralog, pseudogene, or segmental duplication (SMN1, GBA1, CYP2D6, HBA1) or an organelle variant needs calling. Not for general SV calling (sv-cnv-calling) or ACMG classification (variant-interpretation).

• exome-and-panel-analysis

  Use when analyzing WES or targeted panels: capture kit and BED, coverage metrics, per-gene gaps, padding, kit and panel versions, capture CNV. Not for WGS calling (short-read-germline-calling) or SV work (sv-cnv-calling).

• fungal-genomics

  Use when assembling, annotating, or comparing fungal genomes (yeast, molds, plant and human pathogens). Evidence is thin; ask. Not for bacterial genomes (bacterial-genome-assembly-annotation) or amplicon profiling (amplicon-metabarcoding-analysis).

• gene-annotation-te-repeats

  Use when annotating repeats, transposable elements, and genes on a eukaryotic assembly, or transferring an annotation. Not for building the assembly (genome-assembly-and-curation) or choosing a public annotation (reference-genome-selection).

• genome-assembly-and-curation

  Use when assembling a eukaryotic genome from long reads and Hi-C: haplotype strategy, scaffolding, contamination screening, QC, curation. Not for bacteria (bacterial-genome-assembly-annotation) or gene models (gene-annotation-te-repeats).

• genome-browser-review

  Use when visually reviewing variants, SVs, repeats, or splicing in a genome browser, or saving screenshots and review decisions. Not for figure design (plotting) or variant classification (variant-interpretation).

• genomic-prediction

  Use when building, validating, or reporting genomic prediction or polygenic scores in crops, livestock, or humans. Not for general design (study-design) or association testing (association-studies).

• genomics-pitfalls

  Use when code or analysis handles genomic coordinates, variants, builds, sample tables, genotypes, or prediction: checklist of bug classes (coordinates, builds, contigs, REF/ALT, sample order, leakage).

• hpc-resource-planning

  Use when sizing cores, memory, disk, GPU, and runtime for a job, or splitting parallel jobs under the host limit. Not for a slow job that already runs (performance-profiling) or a failed job (troubleshooting-failed-jobs).

• humanizer

  Use when auditing or cleaning text that already exists: measured AI tells, invisible characters, preserved numbers, identifiers, citations. Called by writing. Not for drafting (writing); not a detector-evasion tool.

• lab-rules

  Use at the start of any research task: team standards for reporting style, efficiency, evidence, and briefs. Every agent preloads it.

• lab-team

  Use before delegating to an agent or planning multi-step work: routing rule, agent roster with when not to delegate, and escalation order.

• livestock-genotype-qc-imputation

  Use when cleaning livestock SNP-array genotypes: parentage and pedigree checks, sex and breed checks, imputation to a denser panel. Not for generic array QC (array-genotyping-and-imputation) or the prediction model (genomic-prediction).

• long-read-ont-pacbio

  Use when processing ONT or PacBio reads: basecalling, alignment, small variants, SV, phasing, methylation, assembly, or run metrics. Not for run pass or fail (run-qc-and-demultiplexing) or short-read SV (sv-cnv-calling).

• metagenomics

  Use when analyzing 16S or shotgun microbiome data: host removal, profiling, assembly and binning, contamination controls, compositional statistics, STORMS reporting. Not for isolate genome assembly (bacterial-genome-assembly-annotation).

• methylation-and-epigenomics

  Use when analyzing methylation (human arrays, bisulfite, EM-seq, ONT), ChIP-seq or ATAC-seq peaks, human episignatures, or cell composition, in any organism. Not for expression analysis (rna-seq-and-splicing).

• model-organism-resources

  Use when looking up genes, alleles, phenotypes, orthologs, or GO annotations for a model organism, or citing and downloading these resources. Not for crop or livestock databases (trusted-sources).

• non-human-genomes-and-polyploidy

  Use when working on non-human genomes (crops, trees, livestock, companion, model, wildlife): reference, ploidy, heterozygosity, sex chromosomes, GBS, arrays. Not for prediction (genomic-prediction).

• omics-data-wrangling

  Use when reshaping, filtering, or joining large genomic tables and intervals. Not for coordinate and build bug classes (genomics-pitfalls) or code style (coding-style).

• omics-statistics-and-batch

  Use when testing expression, methylation, or count data, sizing replicates, or detecting and correcting batch effects. Not for the design statement (study-design) or GWAS and QTL (association-studies).

• open-science-reproducible-paper

  Use when you assemble the reproducibility package for a paper: code, environment, workflow, deposits with DOIs, preprint, and journal openness policy. Not for daily logs (analysis-log) or directory layout (data-stewardship).

• orthology-comparative-genomics

  Use when inferring orthogroups, gene-family gain and loss, synteny, or whole-genome duplication across species. Not for tree building (phylogenetic-inference) or selection tests (species-tree-and-selection).

• pangenome-analysis-prokaryote

  Use when comparing gene content across many bacterial or archaeal genomes: core and accessory genes, annotation consistency, recombination. Not for single genomes (bacterial-genome-assembly-annotation) or graph pangenomes (pangenome-and-pav).

• pangenome-and-pav

  Use when building or using a eukaryotic pangenome (graph or linear), calling presence/absence variation, or genotyping SVs across accessions or breeds. Not for bacteria (pangenome-analysis-prokaryote) or single-sample SV calling (sv-cnv-calling).

• pathogen-genomic-epidemiology

  Use when turning pathogen sequence data into consensus genomes, lineages, time trees, and outbreak reports with metadata and sharing. Not for isolate assembly (bacterial-genome-assembly-annotation) or run QC (run-qc-and-demultiplexing).

• performance-profiling

  Use when a step is slower or heavier than expected and you must find the bottleneck and tune threads, I/O, compression, or indexing. Not for sizing a new job (hpc-resource-planning) or a crash (troubleshooting-failed-jobs).

• pharmacogenomics-and-hla

  Use when calling human star alleles, diplotypes, or HLA types from WGS, WES, RNA, or arrays, or reporting drug-gene results. Not for general variant classification (variant-interpretation).

• phylogenetic-inference

  Use when building a maximum-likelihood tree from DNA, codon, or protein sequences of any organism: alignment, trimming, model choice, support, rooting, reporting. Not for gene-tree discordance or selection tests (species-tree-and-selection).

• pipeline-testing-and-ci

  Use when adding tests, regression checks, or continuous integration to a Nextflow, Snakemake, Python, or R pipeline. Not for code style (coding-style) or clinical validation (clinical-pipeline-validation).

• plant-field-trial-gwas

  Use when running GWAS or QTL mapping on replicated, spatially modelled field trials, single or multi-environment. Not for generic association QC (association-studies) or variance components and selection (quantitative-genetics).

• plotting

  Use for any figure, plot, chart, or heatmap: Nature-style 88 mm single-column spec, Okabe-Ito palette, viridis, JPEG 300 dpi by default or PDF on request, templates, and a mandatory legibility check.

• polyploid-genotyping-dosage

  Use when calling or analyzing genotypes in polyploid crops as allele dosage (0 to ploidy). Not for reference choice and assembly (non-human-genomes-and-polyploidy) or the prediction model (genomic-prediction).

• popgen-inbreeding-ne-germplasm

  Use when estimating structure, inbreeding, ROH, or effective population size in non-model, crop, livestock, or genebank samples. Not for human ancestry and admixture workflows (population-genetics) or GWAS (association-studies).

• population-genetics

  Use when computing PCA, ancestry, admixture, kinship, IBD, Fst, or selection scans in any organism, or building a reference panel. Not for GWAS covariates and mixed models (association-studies).

• project-readme

  Use when creating or updating a README.md for a tool, pipeline, analysis project, dataset, or repository: a compact structure that answers what it is, how to install and run it, and how it works.

• proteomics-metabolomics-integration

  Use when adding proteomics or metabolomics to genomic data in any organism: pQTL, colocalization, identifier mapping, normalization, missing values, multi-omics factor models. Not for GWAS itself (association-studies).

• quantitative-genetics

  Use when designing or analyzing breeding data in plants or animals: heritability, BLUP, GBLUP, single-step models, multi-environment trials, G x E, selection response, genetic gain. Not for validation (genomic-prediction).

• rare-disease-prioritization

  Use when ranking human candidate variants in a proband, trio, or family: filter ladder, inheritance models, phenotype ranking, reanalysis. Not for ACMG classification (variant-interpretation) or calling (sequencing-pipelines).

• reduced-representation-genotyping

  Use when designing or analyzing GBS, RAD, DArT, or skim-sequencing data, with or without a reference, including filters and low-coverage imputation. Not for SNP arrays (array-genotyping-and-imputation) or WGS calling (short-read-germline-calling).

• reference-genome-selection

  Use when choosing or recording the reference assembly and annotation for a non-human eukaryote project, or deciding on a new one. Not for liftover mechanics (reference-genomes-and-liftover) or building the assembly (genome-assembly-and-curation).

• reference-genomes-and-liftover

  Use when choosing a reference build or contig naming, or converting coordinates and VCFs between builds. Not for general build bug checks (genomics-pitfalls) or alignment pipelines (sequencing-pipelines).

• repeat-expansions

  Use when testing human samples for pathogenic repeat expansions (FMR1, C9orf72, RFC1, HTT) from short or long reads. Not for SNV, indel, or SV calling (sv-cnv-calling) or paralogs (difficult-genes-and-mito).

• research-ethics

  Use when handling human, animal, or plant-germplasm data or samples, or deciding on consent, privacy, sharing, authorship, or AI-tool use in research and publishing.

• research-planning

  Use when planning a research project, grant, or study in human, animal, or plant genetics: aims, design, analysis plan, resources, risks. Follow the funder template. Not for analysis.

• rna-seq-and-splicing

  Use when analyzing bulk RNA-seq in any organism: strandedness, QC, quantification, differential expression, splicing, fusions, or human rare-disease outliers. Not for assay planning (sample-and-assay-planning).

• run-qc-and-demultiplexing

  Use when a new Illumina or ONT run arrives and needs demultiplexing, FASTQ or run QC, and a pass or re-run call. Not for alignment-level QC (short-read-germline-calling) or sample identity (sample-identity-and-contamination).

• sample-and-assay-planning

  Use when planning sampling, DNA or RNA extraction QC, assay choice, replicates, controls, batch layout, or sample tracking for a genomics study in any organism. Not for analysis.

• sample-identity-and-contamination

  Use when confirming sample sex, relatedness, pedigree, cross-sample contamination, or a suspected swap in sequencing or array data. Not for run-level QC (run-qc-and-demultiplexing) or variant calling (short-read-germline-calling).

• sequencing-pipelines
• short-read-germline-calling
• slides
• somatic-calling
• species-tree-and-selection
• stakeholder-communication
• start-here
• storage-and-data-lifecycle
• study-design
• sv-cnv-calling
• team-playbooks
• translator
• troubleshooting-failed-jobs
• trusted-sources
• update
• variant-interpretation
• workflow-engines-and-containers
• writing

Subagents (14)

• assistant

  Use proactively for cheap volume work: file, log, and code search; lookups by public identifier (rsID, gene, PMID, DOI); job and log checks; running tests; exact mechanical edits. Not for judgment, drafting, or literature synthesis (literature-reviewer).

• bioinformatician

  Use proactively to run one-off sequencing analyses of any organism and long jobs on the user's compute host: QC, alignment, calling, assembly, workflows, sample and genotype tables. Not for reusable tools or packages (software-engineer) or model choice (statistician).

• breeder

  Use proactively for plant or animal breeding questions: breeding design, trials, selection decisions, genetic gain, genomic selection strategy, trait definitions. Not for model fitting (statistician) or code (bioinformatician, software-engineer).

• clinical-geneticist

  Use proactively to classify germline variants (ACMG/AMP, ClinGen) from an evidence table, check gene-disease and phenotype fit, and draft report fields for research use (clinical sign-out stays with an accredited laboratory). Not for single public-ID lookups (assistant) or literature synthesis (lite

• data-steward

  Use proactively when data need a directory layout, metadata, a data dictionary, a FAIR or MIAPPE check, a repository choice, a data management plan, or a consent-scope check before sharing. Not for deciding consent or law questions (the PI decides).

• evolutionary-genomicist

  Use proactively for phylogenetics, species trees, selection tests, comparative genomics, population genomics in non-model species, and model-organism resources. Not for breeding decisions (breeder), outbreak typing (microbial-genomicist), or reusable tools (software-engineer).

• figure-maker

  Use proactively to build publication figures, plots, heatmaps, posters, and slide decks from given data files, with the legibility check. Not for choosing statistics (statistician) or writing captions and text (writer).

• literature-reviewer

  Use proactively to search and synthesize literature on a gene, method, disease, or claim from public identifiers and queries: finds papers, extracts findings, and labels evidence verified or not. Not for single lookups (assistant) or drafting (writer).

• microbial-genomicist

  Use proactively for bacterial, fungal, viral, and pathogen genomics, metagenomics and amplicon work, AMR, and outbreak phylogenetics. Not for human variant work (clinical-geneticist), plant or animal breeding (breeder), or reusable tools (software-engineer).

• reviewer

  Use proactively once before results are reported or shared: independent read-only review of a code diff, manuscript or analysis claims, or a blind variant classification. Not for exploratory work, your own tested edits, or choosing methods (statistician).

• software-engineer

  Use proactively to build or change reusable tools, packages, or libraries in Python, R, C++, Rust, or shell, with tests. Not for one-off analysis runs or long jobs (bioinformatician) or for prose (writer).

• statistician

  Use proactively to choose a statistical method, fit or check models, set multiple-testing and power plans, or recompute reported numbers from the files. Not for the final pre-report review of claims and code (reviewer).

• translator

  Use proactively to translate academic text between American English and another language (default: the user's CLAUDE.md language, else ask): manuscripts, abstracts, captions, letters, finished READMEs. Not for drafting or editing English text (writer).

• writer

  Use proactively to draft or edit manuscripts, abstracts, methods, captions, grants, READMEs, and reports in American academic English with verified citations. Not for translation (translator) or figures and decks (figure-maker).

Similar plugins