Understanding the causes of errors in eukaryotic protein-coding gene prediction: a case study of primate proteomes

Abstract

Background Recent advances in sequencing technologies have led to an explosion in the number of genomes available, but accurate genome annotation remains a major challenge. The prediction of protein-coding genes in eukaryotic genomes is especially problematic, due to their complex exon–intron structures. Even the best eukaryotic gene prediction algorithms can make serious errors that will significantly affect subsequent analyses. Results We first investigated the prevalence of gene prediction errors in a large set of 176,478 proteins from ten primate proteomes available in public databases. Using the well-studied human proteins as a reference, a total of 82,305 potential errors were detected, including 44,001 deletions, 27,289 insertions and 11,015 mismatched segments where part of the correct protein sequence is replaced with an alternative erroneous sequence. We then focused on the mismatched sequence errors that cause particular problems for downstream applications. A detailed characterization allowed us to identify the potential causes for the gene misprediction in approximately half (5446) of these cases. As a proof-of-concept, we also developed a simple method which allowed us to propose improved sequences for 603 primate proteins. Conclusions Gene prediction errors in primate proteomes affect up to 50% of the sequences. Major causes of errors include undetermined genome regions, genome sequencing or assembly issues, and limitations in the models used to represent gene exon–intron structures. Nevertheless, existing genome sequences can still be exploited to improve protein sequence quality. Perspectives of the work include the characterization of other types of gene prediction errors, as well as the development of a more comprehensive algorithm for protein sequence error correction.

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

“…The accurate classification of these features provides the basis for questions focused on species evolution, population dynamics, and functional genomics. Errors in genome annotation are frequent, even among well-studied models, and are propagated through downstream analyses (Deutekom et al, 2019; Meyer et al, 2020; Salzberg, 2019). In most eukaryotes, genome annotation is challenged by partial conservation of sequence patterns, variable lengths of introns, variable distances between genes, alternative splicing, and higher densities of TEs and pseudogenes (Kersey, 2019; Salzberg, 2019).…”

Identifying genes with conserved splicing structure and orthologous isoforms in human, mouse and dog

“…Sequence homology lies at the heart of numerous protein and transcript predictions. However, there is still room for improvement in the underlying comparative genomics and spliced alignment methods [ 22 ]. The latter work shows recurrent challenges in accurately identifying intron-exon boundaries, and in handling non canonical GT and AG splice sites.…”

Functional characterization of prokaryotic dark matter: the road so far and what lies ahead

“…One can thus make the case that homology-based methods are only as good as the databases they rely upon ( Dimonaco et al., 2021 ). Unfortunately, misannotation has not only become commonplace in public databases, as it is also an ongoing problem ( Arkhipova, 2020 , Girardi, Thoden, Holden, 2020 , Impey, Lee, Hawkins, Sutton, Panjikar, Perugini, Soares da Costa, 2020 , Meyer, Scalzitti, Jeannin-Girardon, Collet, Poch, Thompson, 2020 , Nobre, Campos, Lucic-Mercy, Arnholdt-Schmitt, 2016 , Rembeza, Engqvist, 2021 , Schnoes, Brown, Dodevski, Babbitt, 2009 ). To make matters worse, misannotations are seldomly confined to the database where the error first occurred, they often percolate throughout several databases as well ( Promponas et al., 2015 ).…”

“…Previous reports showed that the percentage of incorrect functional assignments in public databases soared from less than 5% in 1998 to as high as 40% in 2005 ( Schnoes et al., 2009 ). More recently, it was reported that up to 50% of protein sequences from public databases contain at least one error ( Meyer et al., 2020 ). As a case in point, some authors express particular concern towards error propagation issuing from draft genome assemblies (e.g., MAGs) ( Arkhipova, 2020 , Koonin, Makarova, Wolf, 2021 , Salzberg, 2019 ).…”