Cross-section

KPV by laboratory

The same compound measured by different laboratories on different gradients. Where the medians differ, the question is whether the material differed or the method did — and for KPV, with a class minimum gradient of 15 min, the method is a live explanation.

69 recordsmedian 99.64 %spec 99.0 %

Laboratory	Reports	Gradient	MS resolution	Median purity %	Published bias (pp)	Bias-adjusted	Method suitability
Janoshik	31	30min	High	99.60	+0.00	99.60	ADEQUATE
PeptideMeter	14	28min	High	99.59	-0.05	99.64	ADEQUATE
Medutest	10	25min	Unit	99.38	+0.08	99.30	ADEQUATE
Nordanalyt	5 n=5	35min	High	99.77	+0.02	99.75	ADEQUATE
Kestrel	5 n=5	15min	Unit	99.93	+0.24	99.69	ADEQUATE
VendorInvestigate	4 n=4	32min	High	99.21	-0.11	99.32	ADEQUATE

Figure. KPV purity by contributing laboratory. Any offset between rows has two possible sources — a real difference in the material each laboratory happened to receive, and the laboratory's own bias — and this figure cannot separate them. The paired analysis that can is on inter-lab agreement.