Computational protocol: Gene Expression of Corals in Response to Macroalgal Competitors

Similar protocols

Protocol publication

[…] Immediately after recording PAM readings, the coral portion that had been in contact with each macroalga or algal mimic was preserved in Trizol (Invitrogen) and frozen. Coral tissue was scraped from the calcium carbonate skeleton where algal contact had occurred and total RNA was extracted according to the manufacturer's protocol. Total RNA was purified using RNeasy MinElute Clean-up kit (Qiagen), and RNA pellets were resuspended in nuclease-free water. RNA concentration and quality were assessed using a NanoDrop ND-1000 spectrophotometer (ThermoScientific). RNA integrity was confirmed using the RNA 6000 Nano kit and Bioanalyzer 2100 (Agilent Technologies).RNA (5 ng) from 3–7 biological replicates for each algal exposure treatment and control, as well as RNA from technical replicates (n = 6) verifying consistency among arrays, was processed and labeled using the One-Color Microarray-Based Gene Expression Analysis kit (Agilent Technologies). Data from technical replicates was not included in the final analysis. Microarray hybridization followed manufacturer's protocol.A custom microarray that included 1,029 and 853 unique coral and Symbiodinium genes, respectively, representing a range of functional pathways, was designed and used to measure changes in gene expression as a result of contact with macroalgae. Genes were acquired from bioinformatic mining of recent transcriptome sequencing projects (www.medinalab.org/zoox/ ) and submissions into public databases (GenBank, www.ncbi.nlm.nih.gov/genbank/). At the time of array design, there were few M. digitata sequences publically available to use during the probe design process. Open reading frames from candidate anthozoan genes were blasted in GenBank to determine probable gene identity with an E-value cutoff of e−6 and to determine that these gene regions were conserved across phylogenetically diverse taxa (typically non-cnidarian invertebrates for coral genes and apicomplexan or plant genes for Symbiodinium). Few genes were restricted to only scleractinian coral species (e.g. scleractinian cysteine-rich peptide genes). Two 60-mer probes for each gene were designed from open reading frames using eArray (Agilent Technologies), and replicated 3–4 times on the microarray in addition to positive (spike-in) and negative controls. A total of 1,185 and 1,061 coral and Symbiodinium genes with some replicated genes with sequences from different species were spotted on the array (). To avoid overrepresentation of single genes in the analysis, expression data from only one homologous gene (the first listed in the probe report) were included in the analysis (direction of expression were the same for the homologous genes, but intensities varied; data not shown). Arrays were scanned using Agilent G2505C Microarray Scanner and Feature Extraction Software 10.7.1.1 (Morehouse School of Medicine, Atlanta, GA). Data were statistically analyzed using JMP Genomics 3 (SAS Institute).Raw spot intensity data were log2 transformed and loess normalized. Background intensity was subtracted from each feature and replicate probes for each gene were averaged. Analysis of variance (ANOVA) was used to detect highly significant expression differences between control, and treatment corals for all pairwise comparisons (P<0.01). By identifying significant differences in gene expression of macroalgal treatments relative to corresponding controls, and not treatment by treatment comparisons, DEGs represented effects from specific treatments and not due to species-specific artifacts (e.g. nucleotide sequence difference between species) unrelated to treatments, or diel cycles of gene expression . Significance levels were adjusted using a false discovery rate correction to control for multiple testing . Hierarchical cluster analysis of significant genes using Ward's method was performed between coral species, macroalgal species and exposure times resulting in clusters based on similarities in gene expression patterns. Principle component analysis (PCA) was conducted to identify patterns of similarity across all experimental variables (coral species, macroalgal type, and exposure time). Analyzing significant gene ontology categories rather than individual genes increases the confidence that a specific biological process is involved in the response to a stimulus since there is evidence that multiple genes are behaving in a consistent, functional manner; therefore, gene functions and ontologies (GO terms) were obtained from the European Bioinformatics Institute (EMBL-EBI) (www.ebi.ac.uk), UniProt (www.uniprot.org), and analyzed using Blast2Go v2.6.4 . […]

Pipeline specifications

Software tools JMP Genomics, Blast2GO
Application RNA-seq analysis