Label-free data standardization for clinical metabolomics

Petr G. Lokhov, Dmitri L. Maslov, Oleg N. Kharibin, Elena E. Balashova, Alexander I. Archakov

In metabolomics, thousands of substances can be detected in a single assay. This capacity motivates the development of metabolomics testing, which is currently a very promising option for improving laboratory diagnostics. However, the simultaneous measurement of an enormous number of substances leads to metabolomics data often representing concentrations only in conditional units, while laboratory diagnostics generally require actual concentrations. To convert metabolomics data to actual concentrations, calibration curves need to be generated for each substance, and this process represents a significant challenge due to the number of substances that are present in the metabolomics data. To overcome this limitation, a label-free standardization algorithm for metabolomics data is required. It was discovered that blood plasma has a set of stable internal standards. The appropriate usage of these newly discovered internal standards provides a background for the label-free standardization of metabolomics data that underlies the SantaOmics (Standardization algorithm for nonlinearly transformed arrays in Omics) algorithm. In this study, using the knee point, it was shown that the metabolomics data can be converted by SantaOmics into a standardized scale that can substitute actual concentration measurements, thus making the metabolomics data directly comparable with each other as well as with reference data presented in the same scale. The developed algorithm sufficiently facilitates the usage of metabolomics data in laboratory diagnostics.