Design and Analysis of Bar-seq Experiments

Abstract

High-throughput quantitative DNA sequencing enables the parallel phenotyping of pools of thousands of mutants. However, the appropriate analytical methods and experimental design that maximize the efficiency of these methods while maintaining statistical power are currently unknown. Here, we have used Bar-seq analysis of the Saccharomyces cerevisiae yeast deletion library to systematically test the effect of experimental design parameters and sequence read depth on experimental results. We present computational methods that efficiently and accurately estimate effect sizes and their statistical significance by adapting existing methods for RNA-seq analysis. Using simulated variation of experimental designs, we found that biological replicates are critical for statistical analysis of Bar-seq data, whereas technical replicates are of less value. By subsampling sequence reads, we found that when using four-fold biological replication, 6 million reads per condition achieved 96% power to detect a two-fold change (or more) at a 5% false discovery rate. Our guidelines for experimental design and computational analysis enables the study of the yeast deletion collection in up to 30 different conditions in a single sequencing lane. These findings are relevant to a variety of pooled genetic screening methods that use high-throughput quantitative DNA sequencing, including Tn-seq.

Plant-derived antifungal agent poacic acid targets β-1,3-glucan

“…The barcode counts for each yeast deletion mutant in the presence of poacic acid were compared to the DMSO control conditions to determine sensitivity or resistance of individual strains (the chemical genetic interaction score) (26). To determine a P value for each top sensitive and resistant mutant, we used the EdgeR package (55,56). A Bonferroni-corrected hypergeometric distribution test was used to search for significant enrichment of GO terms among the top 10 sensitive and resistant deletion mutants (57).…”

Implementation and Data Analysis of Tn-seq, Whole-Genome Resequencing, and Single-Molecule Real-Time Sequencing for Bacterial Genetics

“…The most straightforward method is to apply a Mann-Whitney U test, which makes no assumptions about the underlying distribution of the data (16). Indeed, discrete count data from DNA sequencing reads are not normally distributed; rather, these data follow an overdispersed Poisson (negative binomial) distribution (19). Another method is to perform a log transformation of the ratio and then apply a z-test (6).…”

“…The final method we discuss is the use of RNA sequencing (RNA-seq) analysis software to determine differential mutant abundance levels (19,20). The power of this method of data analysis is the ability to utilize freely available software packages that were developed to identify differences in count data sets-the type of data generated from a Tn-seq experiment (19).…”

“…The power of this method of data analysis is the ability to utilize freely available software packages that were developed to identify differences in count data sets-the type of data generated from a Tn-seq experiment (19). This method, like the ratio of mutant abundance, starts by summing the sequencing read data for each gene (Fig.…”

“…The logic underlying an RNA-seq experiment is conceptually similar to that underlying a Tn-seq experiment: the aim is to identify transcripts or mutants with differential abundance for the treatment and the control. The data for both experiments are discrete count data, to which a negative binomial test may be applied (19). As a result, the tools already generated for RNA-seq that utilize a negative binomial test, such as the EdgeR package (19,21) and DEseq (19,22), have been successfully used to determine differential mutant abundance levels in a barcoded deletion library in Saccharomyces cerevisiae yeast (19).…”