Dried blood spot 3-year stability suitability for biomarker discovery and biobanking initiatives in companion animals

Dried blood spot (DBS) sampling is a rapid, cost-effective, and sustainable method that requires minimal processing. It can be performed using remnant blood samples from clinics or via pinprick at home. The DBS metabolome has similar discriminatory potential to the plasma metabolome, making it suitable for biomarker discovery. This study evaluated the suitability of DBS for long-term storage and its potential for biomarker discovery from companion animal (cats and dogs) biobanks using two different approaches. Objective 1 assessed whether two independent metabolic profiling analyses using duplicate DBS card samples, conducted three years apart would provide similar phenotype discrimination. Objective 2 evaluated the potential to combine independent metabolomics datasets through a bridging sample, facilitating insights from prospective biobanks and enabling the reuse of previously acquired data. Using 54 samples from 5 phenotypic classes (1 cat breed and 4 dog breeds), after three years of storage at −80°C, the DBS metabolome showed a 7% reduction (782–728 metabolites). A thorough investigation indicated that this was likely due to experimental variation rather than time-dependent sample instability. Despite data loss, classification remained consistent across independent datasets using various approaches. Hierarchical cluster analysis (HCA) demonstrated correct pairing of individuals between runs for 81% of individuals. Partial least squares-discriminant analysis and Random Forest (RF) utilized many of the same metabolites to discriminate between classes and RF also accurately predicted blinded samples with similar probability in both runs. RF and HCA analyses of merged data performed well, with RF correctly predicting breed for two blinded samples and HCA correctly pairing 76% of individuals. These findings indicate that quality control samples can be used to normalize independent metabolomics studies to a similar baseline enabling combined analysis at a future time. Overall, the data support DBS as a suitable sample for several years of storage in biobanks, with minimal data loss, making it viable for biomarker discovery. Implementing DBS sampling of remnant blood can accelerate research activities, and merging data through suitable bridging samples can enhance the reuse of metabolome data, providing additional biological and economic value.