Kalkreuter Lab — Extract Dereplication

Match observed masses from crude extracts or fractions against known natural products, cross-checked against the biosynthetic capacity of our own strains.

This is triage, not identification. Mass alone cannot distinguish isomers, and natural products are full of them — treat a match as "worth checking before you chase it," not as an ID. MS/MS, retention time, and ultimately NMR are what actually dereplicate. The value here is narrowing 200 peaks down to the handful worth your time.

Read the cluster percentage, not just the badge. That figure is how much of the reference MIBiG cluster our BGC actually shares. Across our 1,976 BGCs the median is only ~16%, because conserved biosynthetic machinery produces weak hits to reference clusters the strain plainly does not make (74% of ours fall under 30%). Only a match above 70% is treated as real evidence a peak is a known compound; anything lower is shown but not counted against the peak.

Naming the source taxon scores every hit by how closely its reported producing organism matches yours — a peak matching a compound only ever reported from a distant genus is weaker evidence than one reported from your own. Pick a strain when the extract came from a sequenced isolate, which additionally checks whether it carries the cluster. Pick a genus when you know the taxon only from 16S; every genus in the reference set is listed, sequenced or not. If you set both, the strain wins. Narrowing filters hits to that taxon — contaminant and primary-metabolite warnings always come through regardless.

Reference data: 37,074 compounds (7,982 Actinobacterial) from NPAtlas plus the lab's curated Counterstain set, of which 3,114 carry a compound class and mechanism across 301 curated classes. Cross-referenced to MIBiG and to 1,976 BGCs across 48 lab genomes. Taxonomy via NCBI.