Abstract
The Penaeus monodon genome became a subject for extended studies of several aspects of nutrition, growth, and reproduction. In this study, transcriptome from the hepatopancreas and ovary of wild-caught female broodstocks were generated by genome-guided (GG) and de novo (DN) assembly. We compared the effectiveness of these methods in terms of the number of transcripts and their annotations. We analyzed mapping features and differentially expressed genes (DEGs) using three estimation approaches: mapping reads against (i) a genome assembly of P. monodon (reference-based (RB)), transcriptome generated by (ii) GG, and (iii) DN assembly. DN had the highest percentage of mapping rates and annotated aligned reads, leading to 2.09 times more unigenes than GG assembly, with 49% of unigenes matching the blast search, compared to 39.66%. Furthermore, 69% of blasted unigenes from DN assembly were assigned GO terms in DN assembly, compared to 23.9% in GG. Additionally, DEGs identified of the two tissues by DN approach (820) surpassed the total number of DEGs identified by GG (488) and RB (117) approaches. In contrast, the GG approach identified the highest number of DEGs from our genes of interest (93.5%), followed by the DN (82.6%) and the RB (37.3%) approach. The DN assembly is ideal for transcript reconstruction and DEGs recovery, while the GG assembly generated an appropriate database for studying specific genes or sets of genes. We, therefore, recommend using a combination of DN and GG assemblies to improve differential gene expression analysis for non-model organisms with poorly resolved genome annotations.