Comparing de novo and reference-based transcriptome assembly strategies by applying them to the blood-sucking bug Rhodnius prolixus
MetadataShow full item record
High Throughput Sequencing capabilities have made the process of assembling a transcriptome easier, whether or not there is a reference genome. But the quality of a transcriptome assembly must be good enough to capture the most comprehensive catalog of transcripts and their variations, and to carry out further experiments on transcriptomics. There is currently no consensus on which of the many sequencing technologies and assembly tools are the most effective. Many non-model organisms lack a reference genome to guide the transcriptome assembly. One question, therefore, is whether or not a reference-based genome assembly gives better results than de novo assembly. The blood-sucking insect Rhodnius prolixus-a vector for Chagas disease-has a reference genome. It is therefore a good model on which to compare reference-based and de novo transcriptome assemblies. In this study, we compared de novo and reference-based genome assembly strategies using three datasets (454, Illumina, 454 combined with Illumina) and various assembly software. We developed criteria to compare the resulting assemblies: the size distribution and number of transcripts, the proportion of potentially chimeric transcripts, how complete the assembly was (completeness evaluated both through CEGMA software and R. prolixus proteome fraction retrieved). Moreover, we looked for the presence of two chemosensory gene families (Odorant-Binding Proteins and Chemosensory Proteins) to validate the assembly quality. The reference-based assemblies after genome annotation were clearly better than those generated using de novo strategies alone. Reference-based strategies revealed new transcripts, including new isoforms unpredicted by automatic genome annotation. However, a combination of both de novo and reference-based strategies gave the best result, and allowed us to assemble fragmented transcripts.