Technical Series

Biotech companies face a build or buy decision with respect to biostatistics.
This Technical Report discusses the dilemma and shows the advantage of buying a Virtual Biostatistics Consulting Core service.

Software developed by BioRankings based on formal biostatistical theory makes the analysis of microbiome data automatic, unbiased, reproducible, and interpretable. This Technical Report shows how we present microbiome data summarization, perform hypothesis testing comparing the microbiome across groups, calculate sample size and power for microbiome experiments, and estimate the population, or sample-to sample, variability (i.e., population diversity).

Untargeted metabolomics measures metabolites in samples to find those that correlate with subgroups (e.g., disease or healthy tissue). Current analysis pipelines are time consuming and require analysts to make ad hoc subjective decisions. Software developed by our group automates the analysis, removes subjective decision making, and runs fast. Linear models are added which introduces a larger class of biostatistical models and methods for researchers.

This technical report introduces a repeated measures analysis method for the microbiome data using
the generalized Dirichlet-multinomial model. We start by reviewing the concept of compositional data, explain the challenge of the repeated data analysis, present the method, and illustrate its performance in hypothesis testing using simulated data.

If you plan on going to a regulatory agency like the FDA or USDA for your microbiome product, you will need to have biostatistics in order to meet their requirements. This paper shows how to move from exploratory R&D to formally designed experiments to test hypotheses about microbiome data.

This technical report introduces a new approach for analyzing LC/MS untargeted metabolomics data that is automatic and unbiased.

In Technical Report 2 we showed through simulations that all pairwise distances become identical as the number of dimensions approaches infinity. In this Supplement, we demonstrate this theory with real microbiome data.

In biomedical research using high-throughput technology (i.e., -omics), short and wide occurs. It presents huge problemsfrom having a very large number of ways of analyzing the many biological measurements.

In this first Report we show how cluster analysis is highly subjective with results changing for different inputs, present using the Dirichlet multinomial distribution for microbiome data, and finally show an example of this analysis using HMP stool samples.