MetaClean

MetaClean is a web app (with optional desktop CLI) that validates, standardizes, and audits sample metadata for genomics, proteomics, and single-cell projects before it hits the pipeline. It ingests spreadsheets, LIMS exports, and ISA-Tab-like tables, then flags missing fields, inconsistent units, forbidden values, and ontology mismatches (e.g., tissue terms, disease codes). It generates a clean, versioned metadata package with a machine-readable schema, changelog, and export formats for common downstream tools. An AI assistant suggests mappings (e.g., “PBMC” -> CL term), but every change is reviewable and reversible to avoid silent corruption. The goal is boring but critical: fewer pipeline failures, fewer re-runs, and fewer irreproducible datasets caused by sloppy metadata. It’s not a replacement for a LIMS; it’s a quality gate that teams can actually adopt without replatforming.