Biomarker Discovery

The medical community utilizes many metabolites to establish both health and disease such as glucose, ADMA, cholesterol. However, biodiversity is significant from genetics, environmental, life style and microbiota influences that make prognostic and diagnostic markers difficult for common and rare diseases. Our 4-point analysis typically covers 6 to 7 orders of magnitude which allows us to monitor diversity in human populations.  A common question may be, how many patient samples do we need for our discovery of disease progression markers or early diagnostic markers?  Warren Dunn’s paper (Dunn, W.B., Lin, W., Broadhurst, D. et al. Molecular phenotyping of a UK population: defining the human serum metabolome. Metabolomics 11, 9–26 (2015) suggests a minimum of 600 individuals to capture biodiversity.  Others have suggested as little as 300.  This is from random sampling of a healthy or disease population.  Smaller numbers can be effective if the populations are selected by specific criteria and meta data is used to deselect others (high BMI, smokers, those with co-morbidities for example). At HMT we understand biodiversity, look to measure it, and factor that into our analysis with the assumption that any cohort may contain a range of responders and that a healthy cohort will still have metabolites that will range from under 10% variation to over 100% variation.  Our animal and human study analysis starts with sample quality assessment to ensure no sample or sets of samples influences group comparisons or subgroup discovery. Our analysis of data includes subtype assessments, assignment of metabolites as food intake markers, contaminates, microbial and endogenous. No analysis is complete without considering meta data (patient status) and when available, protein expression data of any sort. Typical biomarker panels consist of multiple analyte types (metabolites, proteins, mRNA, meta data).