CFE 2019: Start Registration
View Submission - CMStatistics
B1945
Title: Experimental and computational frameworks for microbiome data analysis assessment Authors:  Hector Corrada Bravo - University of Maryland, College Park (United States) [presenting]
Abstract: Analysis of 16S rRNA marker-gene surveys may be performed by a variety of bioinformatic pipelines and downstream analysis methods. However, appropriate assessment datasets and statistics are needed as there is limited guidance to decide between available analysis methods. Mixtures of environmental samples are useful for assessment as they provide values calculated from measurements of the unmixed samples and the mixture design that can be compared to values recovered by each bioinformatic method. We present an assessment framework for 16S rRNA sequencing analysis methods based on a two-sample titration mixture dataset and metrics to evaluate OTU count table characteristics. The qualitative assessment evaluates feature presence/absence exploiting features only present in unmixed samples or titrations by testing if random sampling can explain their observed relative abundance. The quantitative assessment evaluates how well relative and differential abundance values agree with values expected from the mixture design. To demonstrate the assessment framework, we present results assessing estimates of abundance, differential abundance and beta diversity from count tables generated using three of the most-commonly used bioinformatic pipelines for this analysis. The developed assessment framework serves as a valuable community resource for assessing 16S rRNA marker-gene survey bioinformatic methods.