Pan-transcriptomic analysis identifies coordinated and orthologous functional modules in the diatoms Thalassiosira pseudonana and Phaeodactylum tricornutum

Abstract

Diatoms are important primary producers in the ocean that thrive in diverse and dynamic environments. Their survival and success over changing conditions depend on the complex coordination of gene regulatory processes. Here we present an integrated analysis of all publicly available microarray data for the diatoms Thalassiosira pseudonana and Phaeodactylum tricornutum. This resource includes shared expression patterns, gene functions, and cis-regulatory DNA sequence motifs in each species that are statistically coordinated over many experiments. These data illustrate the coordination of transcriptional responses in diatoms over changing environmental conditions. Responses to silicic acid depletion segregate into multiple distinctly regulated groups of genes, regulation by heat shock transcription factors (HSFs) is implicated in the response to nitrate stress, and distinctly coordinated carbon concentrating, CO₂ and pH-related responses are apparent. Fundamental features of diatom physiology are similarly coordinated between two distantly related diatom species, including the regulation of photosynthesis, cellular growth functions and lipid metabolism. These integrated data and analyses can be explored publicly (http://networks.systemsbiology.net/diatom-portal/).

Introduction

Diatoms are important primary producers in marine ecosystems (Armbrust, 2009, Tsuda et al., 2003), and their ability and capacity to physiologically adjust to changing ocean conditions are critical to their ecological success, both now and in the future. The heterokonts (including diatoms) are phylogenetically distinct from other clades, and represent relatively undiscovered genomic and physiological territory. Recent studies of diatom genomes (Armbrust et al., 2004, Bowler et al., 2008), transcriptomes and proteomes (Allen et al., 2008, Ashworth et al., 2013, Brembu et al., 2011, Carvalho et al., 2011, Chauton et al., 2013, Hook and Osborn, 2012, Kustka et al., 2014, Levitan et al., 2015, Mock et al., 2008, Nymark et al., 2009, Nymark et al., 2013, Sapriel et al., 2009, Shrestha et al., 2012, Thamatrakoln et al., 2012, Thamatrakoln et al., 2013, Valle et al., 2014) hint at complex molecular and regulatory programs that control diatom physiology and acclimation to a variety of different environmental conditions. However, the specific coordination and regulation of these molecular and adaptive processes in diatoms remain largely unknown for multiple reasons: i) the size and complexity of diatom genomes and their molecular responses to change, ii) their genetic uniqueness and lack of sufficient homology in comparison to other well-studied clades, iii) low-throughput experimental genetic approaches, and iv) insufficient data and analysis methods to identify concurrent yet distinct molecular and regulatory pathways operating simultaneously under various conditions.

The integration and analysis of data collected through many different transcriptomic experiments can be used to discover fundamentally coordinated, conditional, and distinct molecular responses that are reflective of gene regulatory processes and cannot be discovered through individual experiments (Brooks et al., 2014, Danziger et al., 2014, Reiss et al., 2006). Data-driven approaches for the discovery of molecular coordination may be particularly useful in the case of diatoms, given the size and novelty of their genomes, transcriptomes and proteomes. To systematically discover and identify modules of putatively coordinated and functionally related diatom genes, we aggregated all available microarray expression data for the model diatoms Thalassiosira pseudonana and Phaeodactylum tricornutum, and performed hierarchical clustering (Eisen et al., 1998) and motif-guided biclustering (Reiss et al., 2006) over many experimental conditions. Highlights of this analysis in the context of specific conditions and functions are discussed herein, and the complete results have been made available for further use and exploration in a web portal at the following url: http://networks.systemsbiology.net/diatom-portal/.