Welcome to the big leaves: Best practices for improving genome annotation in non‐model plant genomes

Abstract

PremiseRobust standards to evaluate quality and completeness are lacking in eukaryotic structural genome annotation, as genome annotation software is developed using model organisms and typically lacks benchmarking to comprehensively evaluate the quality and accuracy of the final predictions. The annotation of plant genomes is particularly challenging due to their large sizes, abundant transposable elements, and variable ploidies. This study investigates the impact of genome quality, complexity, sequence read input, and method on protein‐coding gene predictions.MethodsThe impact of repeat masking, long‐read and short‐read inputs, and de novo and genome‐guided protein evidence was examined in the context of the popular BRAKER and MAKER workflows for five plant genomes. The annotations were benchmarked for structural traits and sequence similarity.ResultsBenchmarks that reflect gene structures, reciprocal similarity search alignments, and mono‐exonic/multi‐exonic gene counts provide a more complete view of annotation accuracy. Transcripts derived from RNA‐read alignments alone are not sufficient for genome annotation. Gene prediction workflows that combine evidence‐based and ab initio approaches are recommended, and a combination of short and long reads can improve genome annotation. Adding protein evidence from de novo assemblies, genome‐guided transcriptome assemblies, or full‐length proteins from OrthoDB generates more putative false positives as implemented in the current workflows. Post‐processing with functional and structural filters is highly recommended.DiscussionWhile the annotation of non‐model plant genomes remains complex, this study provides recommendations for inputs and methodological approaches. We discuss a set of best practices to generate an optimal plant genome annotation and present a more robust set of metrics to evaluate the resulting predictions.

Origin and diversity of Capsella bursa-pastoris from the genomic point of view

“…In order to estimate the quality of annotation, we performed several tests, in particular, calculated BUSCO metrics for the set of predicted proteins (for all proteins together and in a subgenome-wise manner and performed comparison with A. thaliana in terms of such parameters as CDS length, the number of exons, and the percent of single-exon genes. This approach follows recently published recommendations for the improvement and quality control of plant genome annotations [ 21 ]. The BUSCO completeness score for all proteins together was in the range of 97.7–94.1%, for subgenomes—from 90.4 to 96.8% (Additional file 2 : Table S1).…”

Genome sequence of the ornamental plantDigitalis purpureareveals the molecular basis of flower color and morphology variation

“…Similar approaches have been discussed and applied before to reduce large numbers of predicted gene models based on transcription evidence to the most important ones (Liang et al ., 2009; Dohm et al ., 2014; McGrath et al ., 2022). Given that structural annotation workflows do not provide perfectly accurate results, evaluation and filtering is recommended (Vuruputoor et al ., 2023). Although RNA-seq samples from a range of different plant organs and treatments were included in this filtering based on transcript evidence, it is possible that transcripts of bona fide genes were not detected in the RNA-seq data leading to a mis-classification of the corresponding genes as non-expressed pseudogene.…”

Genome Report: The reference genome of an endangered Asteraceae,Deinandra increscenssubsp.villosa, endemic to the Central Coast of California

“…From this, StringTie generated 76,471 transcripts while PsiCLASS produced 116,777. Before filtering, the consensus of transcripts totaled 334 K with 197 K genes, containing many false positives as indicated by the high mono to multi-exonic rate (0.69) (Vuruputoor et al 2023) and low sequence similarity rate within EnTAP (0.65). False positives were well resolved by filtering, with only 69,563 final transcripts and 41,039 genes, with a more typical mono/multi-exonic rate of 0.24 and a high 0.92 sequence similarity rate reported by EnTAP (Table 3).…”