High reproducibility demonstrated in a direct-to-consumer microbiome testing workflow under real-world conditions using fecal reference material

Abstract

Recent studies have documented substantial variation in microbiome profiles recovered from identical biological samples across direct-to-consumer (DTC) gut microbiome testing providers and, in some cases, within individual providers. These reports underscore the need for transparent, published assessments of analytical performance from the companies offering these services. To assess analytical reproducibility in our own testing workflow, we processed 10 replicates of a homogenized National Institute of Standards and Technology (NIST) human fecal reference material (RM 8048) through our full sample processing pipeline, from sample collection to data generation. Replicates were shipped from five U.S. sites over a staggered 35-day window. They were processed amongst over 6,000 customer samples, batched for DNA extraction and sequencing based on their arrival in the lab in nine separate sequencing runs, and analyzed using our standard pipeline. The 10 reference material replicates yielded highly concordant community profiles regardless of shipping time or temperature. Technical variability was minimal (median Bray-Curtis dissimilarity = 0.04) compared with the variation observed in samples from the same Tiny Health customer collected the same day (median Bray-Curtis: 0.08), within one week (0.25), one to two weeks apart (0.28), and between unrelated individuals (0.88). When evaluated using NIST's Coefficient of Disagreement (COD) metric, our workflow yielded a maximum pairwise COD of 0.045, well below NIST's published upper bound of 0.073 for this reference material. All three spike-in control strains were successfully recovered from every sample. These results demonstrate the reproducibility of our DTC microbiome testing workflow under real-world conditions, with robust consistency across sample collection, shipping, laboratory processing, and bioinformatics analysis. They also highlight the value of RM 8048 and similar reference materials for evaluating and transparently reporting analytical performance in DTC microbiome testing.