Functional Analysis of ESTs in Common Wheat
About one million ESTs comprising 125.3 Mb nucleotides were collected from the 51 cDNA libraries constructed from various tissues (including biotic and abiotic tissues) and organs under a range of conditions (Manickavelu et al. 2012). ESTs were assembled with stringent parameters so as to produce 37,138 contigs and 215,199 singlets, 10.6 % of which had no homology with those in the public databases. Using these EST data, we developed the correspondence analysis (CA) method (Yano et al. 2006), as shown in Fig. 10.2. The CA method enables identification and comparisons of significant gene expression with the specific library. This method was applied for comparison of gene expression analysis of susceptive and resistant near-isogenic lines in common wheat infected by Puccinia triticina (Manickavelu et al. 2010). In Fig. 10.2, four libraries developed for powdery mildew and leaf rust (two susceptible and two resistant) have been compared. Using this method, common and specific genes related to treatments or the library were easily selected.
Based on the gene ontology, characteristic proteins were classified according to molecular functions, cellular localization, and biological processes. Furthermore, the unigenes were classified into susceptible and resistant classes based on the EST members assembled from the respective libraries. Several genes showing specific expression in the resistant and susceptible lines could be selected. The molecular pathogenicity of leaf rust after infection in wheat was evaluated, and the EST data were supplied for future studies.
Full-Length cDNA Collection in Common Wheat
For construction of full-length cDNAs, total RNAs were extracted from the 17 tissues studied (Kawaura et al. 2009). Tese RNAs were mixed together to construct a cDNA library with the CAP-trapper method. We used three rounds of cDNA selection. At first, 19,968 clones were supplied for one-path sequencing from both ends. These sequences were clustered and non-redundant groups were selected. In this step, 6,162 sequence-verified full-length cDNAs were obtained after using the primer-walked Sanger sequencing method. Subsequently, data were obtained for an additional 10,645 cDNAs. After the second round, the already sequenced clones were subtracted from the library. An additional 5,712 redundant cDNAs were
Fig. 10.2 Correspondence analysis for characterization of gene expression profiles in four libraries. Four cDNA libraries were developed for powdery mildew and leaf rust (two susceptible and two resistant). Each corner refers to each library; the significant genes related to specific libraries are positioned near, and in the same color as, the corresponding library
sequenced with the 454 FLx + instrument. Finally, 22,519 sequence-verified fulllength cDNAs are now available. These FLcDNAs showed a wide size distribution with a range of 64–8,983 bp, and the mean size of these FLcDNAs was 1,848 bp larger than those of barley (1,711 bp; Matsumoto et al. 2011) and rice (1,746 bp; Kikuchi et al. 2003).
We performed RNAseq analysis of diploid ancestors to assign these FLcDNAs to three genomes, namely A, B, and D. Total RNAs were extracted from seedlings, roots and spikes of Triticum urartu Tum., Aegilops speltoides Tausch, and Aegilops tauschii Coss. These RNAs were utilized for RNAseq by using the Illumina Hiseq 2000 system. For each RNA sample, more than 50 Gb were read. These RNA data were grouped to generate contig clusters for each species. In this case, we used data for 16,807 FLcDNAs. These FLcDNA data were compared against contigs from each species. Top hits were selected based on homology and SNP patterns amongst genes from the three genomes. Thus, 4,759 cDNAs were assigned to the A genome, 2,849 to B, and 5,343 to the D genome. In total, 77 % of genes could be assigned. The remaining 23 % had no counterpart in the diploid RNAseq data pool. The B genome had a lesser number of expressed genes than the other genomes. This finding suggests that the lesser sequence homology of genes between the B genome of common wheat and the S genome of Ae. speltoides might disturb, to some extent, the ability to assign groups.