How would you troubleshoot an assay that suddenly starts producing highly variable baseline noise across replicates?
