BackgroundTumourous stem mustard (Brassica juncea var. tumida Tsen et Lee) is an economically and nutritionally important vegetable crop of the Cruciferae family that also provides the raw material for Fuling mustard. The genetics breeding, physiology, biochemistry and classification of mustards have been extensively studied, but little information is available on tumourous stem mustard at the molecular level. To gain greater insight into the molecular mechanisms underlying stem swelling in this vegetable and to provide additional information for molecular research and breeding, we sequenced the transcriptome of tumourous stem mustard at various stem developmental stages and compared it with that of a mutant variety lacking swollen stems.ResultsUsing Illumina short-read technology with a tag-based digital gene expression (DGE) system, we performed de novo transcriptome assembly and gene expression analysis. In our analysis, we assembled genetic information for tumourous stem mustard at various stem developmental stages. In addition, we constructed five DGE libraries, which covered the strains Yong'an and Dayejie at various development stages. Illumina sequencing identified 146,265 unigenes, including 11,245 clusters and 135,020 singletons. The unigenes were subjected to a BLAST search and annotated using the GO and KO databases. We also compared the gene expression profiles of three swollen stem samples with those of two non-swollen stem samples. A total of 1,042 genes with significantly different expression levels occurring simultaneously in the six comparison groups were screened out. Finally, the altered expression levels of a number of randomly selected genes were confirmed by quantitative real-time PCR.ConclusionsOur data provide comprehensive gene expression information at the transcriptional level and the first insight into the understanding of the molecular mechanisms and regulatory pathways of stem swelling and development in this plant, and will help define new mechanisms of stem development in non-model plant organisms.

BackgroundNucleotide binding site (NBS) genes encode a large family of disease resistance (R) proteins in plants. The availability of genomic data of the two diploid cotton species, Gossypium arboreum and Gossypium raimondii, and the two allotetraploid cotton species, Gossypium hirsutum (TM-1) and Gossypium barbadense allow for a more comprehensive and systematic comparative study of NBS-encoding genes to elucidate the mechanisms of cotton disease resistance.ResultsBased on the genome assembly data, 246, 365, 588 and 682 NBS-encoding genes were identified in G. arboreum, G. raimondii, G. hirsutum and G. barbadense, respectively. The distribution of NBS-encoding genes among the chromosomes was nonrandom and uneven, and was tended to form clusters. Gene structure analysis showed that G. arboreum and G. hirsutum possessed a greater proportion of CN, CNL, and N genes and a lower proportion of NL, TN and TNL genes compared to that of G. raimondii and G. barbadense, while the percentages of RN and RNL genes remained relatively unchanged. The percentage changes among them were largest for TNL genes, about 7 times. Exon statistics showed that the average exon numbers per NBS gene in G. raimondii and G. barbadense were all greater than that in G. arboretum and G. hirsutum. Phylogenetic analysis revealed that the TIR-NBS genes of G. barbadense were closely related with that of G. raimondii. Sequence similarity analysis showed that diploid cotton G. arboreum possessed a larger proportion of NBS-encoding genes similar to that of allotetraploid cotton G. hirsutum, while diploid G. raimondii possessed a larger proportion of NBS-encoding genes similar to that of allotetraploid cotton G. barbadense. The synteny analysis showed that more NBS genes in G. raimondii and G. arboreum were syntenic with that in G. barbadense and G. hirsutum, respectively.ConclusionsThe structural architectures, amino acid sequence similarities and synteny of NBS-encoding genes between G. arboreum and G. hirsutum, and between G. raimondii and G. barbadense were the highest among comparisons between the diploid and allotetraploid genomes, indicating that G. hirsutum inherited more NBS-encoding genes from G. arboreum, while G. barbadense inherited more NBS-encoding genes from G. raimondii. This asymmetric evolution of NBS-encoding genes may help to explain why G. raimondii and G. barbadense are more resistant to Verticillium wilt, whereas G. arboreum and G. hirsutum are more susceptible to Verticillium wilt. The disease resistances of the allotetraploid cotton were related to their NBS-encoding genes especially in regard from which diploid progenitor they were derived, and the TNL genes may have a significant role in disease resistance to Verticillium wilt in G. raimondii and G. barbadense.Electronic supplementary materialThe online version of this article (doi:10.1186/s12864-017-3682-x) contains supplementary material, which is available to authorized users.

Circular RNAs (circRNAs), a class of recently discovered non-coding RNAs, play a role in biological and developmental processes. A recent study showed that circRNAs exist in plants and play a role in their environmental stress responses. However, cotton circRNAs and their role in Verticillium wilt response have not been identified up to now. In this study, two CSSLs (chromosome segment substitution lines) of G.barbadense introgressed into G. hirsutum, CSSL-1 and CSSL-4 (a resistant line and a susceptible line to Verticillium wilt, respectively), were inoculated with V. dahliae for RNA-seq library construction and circRNA analysis. A total of 686 novel circRNAs were identified. CSSL-1 and CSSL-4 had similar numbers of circRNAs and shared many circRNAs in common. However, CSSL-4 differentially expressed approximately twice as many circRNAs as CSSL-1, and the differential expression levels of the common circRNAs were generally higher in CSSL-1 than in CSSL-4. Moreover, two C-RRI comparisons, C-RRI-vs-C-RRM and C-RRI-vs-C-RSI, possessed a large proportion (approximately 50%) of the commonly and differentially expressed circRNAs. These results indicate that the differentially expressed circRNAs may play roles in the Verticillium wilt response in cotton. A total of 280 differentially expressed circRNAs were identified. A Gene Ontology analysis showed that most of the ‘stimulus response’ term source genes were NBS family genes, of which most were the source genes from the differentially expressed circRNAs, indicating that NBS genes may play a role in Verticillium wilt resistance and might be regulated by circRNAs in the disease-resistance process in cotton.