BatchQC

BatchQC is a web app (with optional desktop CLI) that detects, quantifies, and explains batch effects in omics datasets (RNA-seq, scRNA-seq, proteomics, metabolomics) before downstream analysis. Users upload a count matrix plus sample metadata, and the app runs standardized diagnostics: PCA/UMAP colored by batch covariates, variance partitioning, replicate concordance, and automated “suspicious covariate” discovery. It then proposes realistic mitigation steps (e.g., re-randomization checks, covariate inclusion, ComBat/limma removal, Harmony/scVI options) and generates a shareable QC report for collaborators and reviewers. This is an AI + traditional app: AI summarizes findings, flags likely confounders, and drafts methods text, but core statistics remain deterministic and auditable. The honest value: it won’t magically fix bad experimental design, but it will surface problems early and make decisions defensible.