Lipid concentrations in standard reference material (SRM) 1950: results from an interlaboratory comparison exercise for lipidomics

Abstract

Lipid consensus estimates and associated uncertainties were determined using three methods, the Vangel-Rukhin (VR), the DerSimonian-Laird (DSL), and the median of means (MEDM). The VR method (8, 9) is a weighted mean based on maximum likelihood. The derivation and mathematical details of the VR mean and its associated standard uncertainty are given in a previous publication This method incorporates both inter-laboratory and intra-laboratory variances in determining the weights. Maximum likelihood approaches have excellent statistical properties; however, the VR method is not robust against extreme outliers. In these cases, the VR weighting method gives the outlying laboratory value greater influence and results in unrepresentative estimates (i.e., the resulting value reflects neither the non-outlying values or the outlying value). Thus, to obtain a reasonable mean using this approach, the extreme outlying lipid values would need to be omitted from the estimation. The VR means were calculated with no outliers omitted.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 15 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.