miniAgent

Manage the BioAPEX current-feature workflow from scoping through review and completion

Turn an analysis request into a Slurm-ready execution plan with commands, resource assumptions, and job structure.

Scale a buffer recipe to a target volume and compute component masses/volumes.

Save a fetched summary or document to the knowledge base for later retrieval (e.g. after PubMed/UniProt lookup).

Interpret scRNA clusters using marker genes and suggest cell type or state.

Critically evaluate a perturbation hypothesis — challenge assumptions, propose negative controls, and flag confounders.

Explain where project data and outputs live (e.g. GPFS, data/, predictions/) from knowledge base.

Interpret a differential expression result with replicate-aware context, likely confounders, and the next analysis decision.

Calculate C1V1=C2V2 dilutions, serial dilutions with plate-based replicates, and budget-conscious volume planning that accounts for micropipette systematic error.

Advise on doublet detection and removal for scRNA (scrublet, scDblFinder, best practices).

Get a short summary of gene function from UniProt/NCBI Gene and related pathways.

Run or interpret gene set enrichment (Enrichr/g:Profiler) and summarize results.

Normalize gene symbols, aliases, and species context before downstream lookup or analysis.

Generate candidate perturbation hypotheses (targets, mechanism, expected signature, experiment plan) for a given biological context and phenotype goal.

Precheck biological and interpretation risks before selecting perturbation targets or guide designs.

Summarize the current literature consensus, disagreements, and evidence strength for a biological question.

Validate whether a marker set supports a proposed cell type or cell state and explain the main caveats.

Convert between molarity, mass, and volume using molecular weight.

Look up gene description and location via NCBI Gene (esearch/efetch).

Advise on normalization for scRNA (library size, log, SCTransform, etc.).

Triage a paper for biological relevance, extract the main claims, and separate evidence-backed takeaways from abstract-only impressions.

Look up pathway or GO/Reactome/KEGG term definitions and key genes.

Compare gene signatures across perturbations and suggest clustering or grouping.

Generate a 96- or 384-well plate map for perturbations and replicates.

Advise on PCR primer design (Tm, length, GC%, avoiding secondary structure).

Retrieve a lab protocol or SOP from the local knowledge base and summarize the parts that matter for the current experiment.

Fetch and summarize a PubMed article by PMID (abstract, key methods, key findings).

Search PubMed for articles by query; return top N PMIDs with titles, year, journal, and link.

Parse common pipeline log/output and summarize metrics and status.

Propose safe commands and directory hygiene for running a pipeline.

Suggest Scanpy/Python commands for basic scRNA QC (filtering, mito, UMI).

Turn single-cell RNA-seq summary metrics into a practical QC checklist with explicit assumptions and threshold recommendations.

Generate an sbatch script template for GPU or CPU jobs (Slurm).

Rank perturbation targets by drugability, pathway relevance, and safety given constraints.

Fetch protein entry from UniProt by ID or gene name (function, domains, organism).

Convert between common lab units (mg/mL to M, % to molar, ng/µL, etc.) given MW where needed.

Triage whether ambient RNA contamination is a plausible explanation for suspicious single-cell expression patterns.

Recommend a batch-integration strategy and explain the tradeoffs for preserving biology versus removing technical effects.

Count the number of lines in a given file

Build an evidence matrix linking a gene list to a phenotype, pathway, or biological question using literature and local knowledge.