Comparative performance of transcriptome assembly programs for saffron (Crocus sativus L.)

Saffron (Crocus sativus L.) belonging to Iridaceae family as a source of apocarotenoids is one of the most valuable spices and medicinal plants in the world. Because of the large size and high complexity of saffron genome, its sequencing remains a challenge. The arrival of next-generation sequencing technologies (NGS) has allowed the rapid and efficient development for RNA sequencing. De novo assembly of transcriptome from short-read RNA-Seq data provides a great resource for the study of species without a reference genome. De novo assembly of the transcriptome has some unique challenges, particularly in the case of plants, which possess a large amount of paralogs, orthologs, homoeologs and isoforms. In this research, we attempted to compare the performance of de novo assembly tools including BinPacker, Bridger, Oases-Velvet and Trinity through consideration of a quality metrics such as N50 length, the total number of contigs and alignment scores. The results of these analyses revealed that assembly using Bridger had a superior performance for saffron transcriptome, Oases suffered from relatively high chimera rates and redundancies which causes genes family with high similarity assembled into one transcript, Trinity performs worse than Bridger in the increase of false positives. Our comparison study will assist researchers in selecting a well-suited assembler and offer essential information for the improvement of existing assemblers.