Selecting Superior <i>De Novo</i> Transcriptome Assemblies: Lessons Learned by Leveraging the Best Plant Genome

<div><p>Whereas <i>de novo</i> assemblies of RNA-Seq data are being published for a growing number of species across the tree of life, there are currently no broadly accepted methods for evaluating such assemblies. Here we present a detailed comparison of 99 transcriptome assemblies, generated with 6 <i>de novo</i> assemblers including CLC, Trinity, SOAP, Oases, ABySS and NextGENe. Controlled analyses of <i>de novo</i> assemblies for <i>Arabidopsis thaliana</i> and <i>Oryza sativa</i> transcriptomes provide new insights into the strengths and limitations of transcriptome assembly strategies. We find that the leading assemblers generate reassuringly accurate assemblies for the majority of transcripts. At the same time, we find a propensity for assemblers to fail to fully assemble highly expressed genes. Surprisingly, the instance of true chimeric assemblies is very low for all assemblers. Normalized libraries are reduced in highly abundant transcripts, but they also lack 1000s of low abundance transcripts. We conclude that the quality of <i>de novo</i> transcriptome assemblies is best assessed through consideration of a <i>combination</i> of metrics: 1) proportion of reads mapping to an assembly 2) recovery of conserved, widely expressed genes, 3) N<sub>50</sub> length statistics, and 4) the total number of unigenes. We provide benchmark Illumina transcriptome data and introduce <i>SCE</i>RNA, a broadly applicable modular protocol for <i>de novo</i> assembly improvement. Finally, our <i>de novo</i> assembly of the <i>Arabidopsis</i> leaf transcriptome revealed ~20 putative <i>Arabidopsis</i> genes lacking in the current annotation.</p></div>