EAGER: EviAnn, a Novel Genome Annotation Software for Plant Genomes

Johns Hopkins University, Baltimore MD

Investigators

Abstract

Genome annotation has always been a central aspect of genome projects, and its importance has only grown in the wake of the past decade's rapid advances in genome sequencing and assembly. Annotation locates functional elements on a genome, including protein-coding genes, noncoding RNA genes, repeat elements, and other genomic features. Identification of protein-coding genes is generally considered to be the most important component of annotation, although the need for more information about noncoding RNAs is increasing rapidly. Large-scale quantitative genetic analyses such as Genome Wide Association Studies (GWAS) usually focus on mutations associated with protein-coding genes, including mutations that can disrupt a protein by introducing frameshifts or premature stop codons. If a gene is missing or incorrectly annotated, these analyses return spurious results or fail entirely. For these reasons, it is vitally important to correctly locate all protein-coding sequences on a genome. All leading annotation methods today, including the widely used MAKER and BRAKER systems, rely critically on ab initio gene finding programs that use gene structure and known signature motifs to detect genes. However, based on experimental observations with modern transcriptome sequencing data, this strategy seems to be no longer adequate. The goal of this EAGER project is to develop a radically different approach to genome annotation that relies mostly on experimental evidence, avoiding the use of ab initio gene finding techniques to locate functional elements on the genome. Software developed as well as instructional videos will be readily available through GitHub and YouTube, respectively. If successful, it is anticipated that EviAnn will be useful to any groups that are working on assembly and annotation of plant genomes, thus enabling both basic and applied research needed for advancing crop improvement and the U.S. bioeconomy. This project will develop an open-source software package that is based on processing and cross-referencing different types of evidence, resulting in a faster and more transparent annotation process, where one can trace the origin of every annotated gene or transcript back to transcriptional evidence, to protein alignment evidence, or both. Specifically, the EviAnn (Evidence-based Annotation) pipeline, together with EviProt, a novel protein-to-genome alignment package, will produce genome annotation by combining transcript assemblies created from Illumina RNA-seq data, PacBio IsoSeq and Oxford Nanopore RNA/cDNA sequencing data, alignments of transcripts from closely related species (if available), as well as alignments of proteins from at least one other related species. If successful, the software developed will likely be widely adopted by the community because of its expected superior performance and ease of use, resulting in better gene annotation that will enable advances in plant breeding, discovery of genes for disease and pest resistance, and ultimately will advance the frontiers of plant genomic research. To further promote wide adoption of EviAnn, the project will produce a series of educational videos that describes basic usage of the software by walking through all the steps using examples from small genomes. For example, the videos will show users will demonstrate how to deal with common failures in installing and running the software and how to search for transcript and protein data for related species in public databases. All project outcomes will be made available to research communities through deposition at and dissemination through public repositories. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.