The importance of study design for detecting differentially 1 abundant features in high - throughput experiments

8 The use of high-throughput experiments, such as RNA-seq, to simultaneously identify 9 differentially abundant entities across conditions has become widespread, but the systematic 10 planning of such studies is currently hampered by the lack of general-purpose tools to do so. 11 Here we demonstrate that there is substantial variability in performance across statistical 12 tests, normalization techniques and study conditions, potentially leading to significant 13 wastage of resources and/or missing information in the absence of careful study design. We 14 present a broadly applicable experimental design tool called EDDA, and the first for single15 cell RNA-seq, Nanostring and Metagenomic studies, that can be used to i) rationally choose 16 from a panel of statistical tests, ii) measure expected performance for a study and iii) plan 17 experiments to minimize mis-utilization of valuable resources. Using case studies from recent 18 single-cell RNA-seq, Nanostring and Metagenomics studies, we highlight its general utility 19 and, in particular, show a) the ability to correctly model single-cell RNA-seq data and do 20 comparisons with 1/5 the amount of sequencing currently used and b) that the selection of 21 suitable statistical tests strongly impacts the ability to detect biomarkers in Metagenomic 22 studies. Furthermore, we demonstrate that a novel mode-based normalization employed in 23 EDDA uniformly improves in robustness over existing approaches (10-20%) and increases 24 precision to detect differential abundance by up to 140%. 25