Annotation of the Triticum dicoccoides Genome Provides Insights into Genome Organization and Cereal Evolution

The tetraploid wild Emmer, Triticum dicoccoides, is an important ancestor of most domesticated wheat accessions. An annotation pipeline was set up to predict gene structures on a Triticum dicoccoides genome assembly. The annotation pipeline combined information from gene structures of closely related grass species and gene expression data. We used Genomethreader to perform splice aware alignments between protein sequences as well as cDNA sequences from related grass species on the Triticum dicoccoides assembly after masking for repetitive regions. RNA-seq reads from 20 different tissues were aligned against the genome assembly using Tophat2 and assembled by using Cufflinks. Results from the reference based approach were merged with gene structure information from RNA-seq experiment by using Cuffcompare and redundant transcripts were removed from the results. Potential transcript sequences were extracted and the putative ORFs were determined with the Transdecoder software. Based on PFAM signatures and peptide homology searches against a comprehensive high-confidence plant protein data set including Arabidopsis thaliana, Zea mays, Brachypodium distachyon, Oriza sativa, and Sorghum bicolor sequences, the most reliable peptide translations were selected for all predicted transcripts. The analysis resulted in structural information of more than 71,000 high confidence genes and more than 300,000 transcripts. Functional annotations (“human readable descriptions”) were computed for all high confidence genes using the AHRD (MPIZ Cologne) pipeline, including gene onthology (GO) terms and protein domains (PFAM/INTERPRO).