Detecting gene ontology misannotations using taxon-specific rate ratio comparisons
Year: 2020
Type: other
Abstract: INTRODUCTION Many protein function databases are built on automated or semi-automated Gene Ontology (GO) curations and can contain various annotation errors (Andorf, et al., 2007; Schnoes, et al., 2009; Skunca, et al., 2012). The correction of such misannotations is critical to improving the accuracy and reliability of the databases. METHODS We proposed a new approach to detect potentially incorrect GO annotations by comparing the ratio of annotation rates (RAR) for the same GO term across different taxonomic groups, where those with a relatively low RAR usually correspond to incorrect annotations. RESULTS As an illustration, we applied the approach to 20 commonly-studied species in two recent UniProt-GOA (Huntley, et al., 2015) releases and identified 250 potential misannotations in the 2018-11-6 release, where only 25% of them were corrected in the 2019-6-3 release by UniProt and other members of the GO consortium. Importantly, 56% of the misannotations are “Inferred from Biological aspect of Ancestor (IBA)” which is in contrast to previous observations (Jones, et al., 2007) that attributed misannotations mainly to “Inferred from Sequence or structural Similarity (ISS)”, probably reflecting an error source shift due to the new developments of function annotation databases (Gaudet, et al., 2011). Typical IBA misannotations include assignment of “cell wall” to animal proteins, and association of “mitochondrion” and “nucleus” to bacterial proteins. Detailed analyses show that these misannotations may have been caused by over-interpretation of phylogenetic evidences during their semi-automated curation. The frequent instances of IBA misannotation are consistent with our separate estimation of misannotation rates in UniProt-GOA using all 10,523 unique annotations from release 2018-11-6 that are later experimentally confirmed or rejected (with a “NOT” qualifier) in release 2019-6-3. This systematic estimation suggests that IBA terms have the highest error rate (31%), followed by fully automatically annotated “Inferred from Electronic Annotation (IEA)” terms (5%), while experimental annotations are almost error-free. These results demonstrated a simple but efficient misannotation detection approach that is useful for development of new taxon constraints (Deegan, et al., 2010) in large-scale comparative protein function studies. Our identified misannotations are being reconsidered by UniProt, neXtProt and the GO consortium for correction in a future release. Additionally, the neXtProt database is considering downgrading IBA annotations for human protein from “Gold” to “Silver” status to reflect their lower quality. AVAILABILITY The source code for performing the RAR analysis and the identified misannotations are available at https://zhanglab.ccmb.med.umich.edu/RAR REFERENCES Andorf, C., Dobbs, D. and Honavar, V. Exploring inconsistencies in genome-wide protein function annotations: a machine learning approach. Bmc Bioinformatics 2007;8. Deegan, J.I., Dimmer, E.C. and Mungall, C.J. Formalization of taxon-based constraints to detect inconsistencies in annotation and ontology development. Bmc Bioinformatics 2010;11. Gaudet, P. , et al. Phylogenetic-based propagation of functional annotations within the Gene Ontology consortium. Brief Bioinform 2011;12(5):449-462. Huntley, R.P. , et al. The GOA database: Gene Ontology annotation updates for 2015. Nucleic Acids Research 2015;43(D1):D1057-D1063. Jones, C.E., Brown, A.L. and Baumann, U. Estimating the annotation error rate of curated GO database sequence annotations. Bmc Bioinformatics 2007;8. Schnoes, A.M. , et al. Annotation Error in Public Databases: Misannotation of Molecular Function in Enzyme Superfamilies. Plos Comput Biol 2009;5(12). Skunca, N., Altenhoff, A. and Dessimoz, C. Quality of Computationally Inferred Gene Ontology Annotations. Plos Comput Biol 2012;8(5).
Institution: University of Michigan
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: gold
OpenAlex ID: W3046881814