Making Sense of Transcriptome Assemblies

Here we describe a framework for the analysis, visualization, and annotation of de novo assembled transcriptomes.  The mRNA-Seq libraries were constructed from two developmental conditions of Pachycereus pringlei roots. For the de novo assembly, we used 14 million of 85 nt long Illumina reads. De novo assembly and all consecutive analysis steps were performed in CLC Genomics Workbench. The assembly produced 27,317 contigs with the longest assembled transcript over 15 kb, as expected for a plant species. We compared the expression patterns in two developmental conditions by RNA-Seq mapping and counting. The reads from both libraries were mapped separately to the entire transcriptome assembly.  The differentially expressed (up, and down regulated) contigs were extracted from the assembly and submitted to the BlastX analysis against all dicot RefSeq proteins. The output was further analyzed using the Blast2GO plugin in the Genomics Workbench. This step allowed us to mine, annotate, and visualize the major functional, biological, and cellular components in both datasets. We also constructed the k-mer based phylogenetic trees from the differentially expressed transcripts, and added the produced functional annotations to the trees. This step combined the functional and phylogenetic analysis, and allowed us to visualize and extract particular functional gene categories from the phylogenetic trees. The extracted sets of genes of interest were translated to proteins, and submitted to Pfam domain search and transmembrane helix prediction. The described framework allowed for global analysis of the assemblies, and also for the detailed functional characterization of particular genes.