A comparative study of metatranscriptomic assessment methods to characterize Microcystis blooms. (8th November 2021)

• Record Type: Journal Article
• Authors: Pound, Helena L.; Gann, Eric R.; Wilhelm, Steven W.
• Is Part Of: Limnology and oceanography, methods. Volume 19:Number 12(2021)
• Page Start: 846
• Page End: 854
• Publication Date: 2021-11-08
• Subjects: Limnology -- Methodology -- Periodicals; Oceanography -- Methodology -- Periodicals; 551.48
• DOI: 10.1002/lom3.10465
• Languages: English
• ISSNs: 1541-5856

Abstract

Abstract: Harmful algal blooms are increasing in duration and severity globally, resulting in increased research interest. The use of genetic sequencing technologies has provided a wealth of opportunity to advance knowledge, but also poses a risk to that knowledge if handled incorrectly. The vast numbers of sequence processing tools and protocols provide a method to test nearly every hypothesis, but each method has inherent strengths and weaknesses. Here, we tested six methods to classify and quantify metatranscriptomic activity from a harmful algal bloom dominated by Microcystis spp. Three online tools were evaluated (Kaiju, MG‐RAST, and GhostKOALA) in addition to three local tools that included a command line BLASTx approach, recruitment of reads to individual Microcystis genomes, and recruitment to a combined Microcystis composite genome generated from sequenced isolates with complete, closed genomes. Based on the analysis of each tool presented in this study, two recommendations are made that are dependent on the hypothesis to be tested. For researchers only interested in the function and physiology of Microcystis spp., read recruitments to the composite genome, referred to as "Frankenstein's Microcystis, " provided high total estimates of transcript expression. However, for researchers interested in the entire bloom microbiome, the online GhostKOALA annotation tool, followed by subsequent read recruitments, provided functional and taxonomic characterization, in addition to transcript expression estimates. This study highlights the critical need for careful evaluation of methods before data analysis.