NGS profiling of invertebrate immune systems

Abstract

Much of what is known about invertebrate immunity has been discovered through genome-wide studies of organisms like Drosophila with well-annotated and curated genomes and transcriptomes. The advanced state of annotation and curation allowed for interrogation of target sequences with a diverse array of "omic" approaches to characterize immune pathways and responses. Recent developments in next-generation sequencing (NGS) approaches, like RNA-Seq, have expanded the scope of immune studies to both model and non-model organisms by allowing differentially expressed (DE) transcripts to be identified without prior knowledge of target sequences. My thesis focuses on innate immunity in invertebrates using RNA-Seq to profile immune responses, immune priming, and trans-generational immune-priming (TGIP). I used RNA-Seq to explore TGIP in the tobacco hornworm Manduca sexta and the interaction of symbiotic state and immune challenge in the symbiotic anemone Aiptasia pallida. In the course of performing bioinformatic analyses of these RNA-Seq datasets, it became clear that annotation was one of the most difficult and time-consuming steps of RNA-Seq analyses, and to make this process more straightforward for future projects I integrated the R code developed in the first two chapters into an Amazon Web Services (AWS) - enabled Bioconductor annotation package called Trans2Kegg. In my first chapter, I examined trans-generational gene expression patterns in immune-challenged Manduca sexta and their offspring. Maternal exposure to both live and heat-killed Serratia marcescens resulted in strong and significant trans-generational impacts on gene expression patterns in their offspring, and within the closely interrelated immune system and signal transduction categories, I identified a total of 27 highly differentially expressed (DE) genes (|log base 2 fold change| >1) expressed trans-generationally due to Serratia exposure. My results indicate that immune priming includes up-regulation of genes associated with pathogen recognition, up-regulation of coagulation genes associated with encapsulating pathogens, and up-regulation of general fitness and reproductive health genes. In my second chapter, I explored gene expression patterns in symbiotic and aposymbiotic Aiptasia pallida to determine the effect of symbiosis on the anemone's immune response. Multivariate and univariate analyses of Aiptasia gene expression demonstrated that exposure to live Vibrio coralliilyticus and menthol bleaching had strong and significant impacts on transcriptome-wide gene expression that are independent or additive in their effects, but not interactive. Vibrio exposure had the strongest impact (4,164 DE genes) followed by menthol treatment (1,114 DE genes) and then the additive combinations of Vibrio and menthol (472 DE genes). KEGG enrichment analyses identified 11 pathways - involved in immunity (5), transport and catabolism (4) and cell growth and death (2) - that were enriched due to both Vibrio and/or menthol exposure. Immune pathways showing strongest differential expression included complement, coagulation, NOD, and Toll for Vibrio exposure and coagulation and apoptosis for menthol. For my third chapter, I developed the Bioconductor package Trans2Kegg to facilitate the annotation of differentially expressed (DE) genes. CPU-intensive processes run on either NCBI BLAST servers or the NCBI BLAST AMI hosted on AWS, allowing annotation to be run on a standard desktop or laptop computer. The annotation process requires only two inputs - a DESeqDataSet and the FASTA file to which RNA-Seq reads were aligned. The minimal inputs make Trans2Kegg annotation straightforward with basic R skills. Running BLAST on AWS, Trans2Kegg can annotate a thousand genes to KEGG orthologs, pathways, pathway classes, and pathway categories in less than eight hours while incurring minimal (less than twenty dollars) AWS charges. The AWS option reduces annotation time 14-fold compared to BLAST on NCBI servers. Trans2Kegg uses a reciprocal best match approach to eliminate duplicate mappings of transcripts to orthologs, and by mapping to species-independent KEGG ortholog identifiers allows for cross-species comparisons of DE results.