Chimeric mis-annotations of genes remain pervasive in eukaryotic non-model organisms

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Accurate protein-coding gene annotation is essential for non-model organisms.
• Chimeric gene mis-annotations, where adjacent genes fuse incorrectly, are a significant problem.
• Annotation inertia propagates these errors, complicating downstream genomic analyses.

Purpose of the Study:

• To investigate the prevalence of chimeric mis-annotations in diverse genomes.
• To evaluate the effectiveness of machine-learning tools in identifying these errors.

Main Methods:

• Analysis of 30 recently annotated genomes (invertebrates, vertebrates, plants).
• Identification and confirmation of chimeric gene mis-annotations.
• Utilizing machine-learning annotation tools (e.g., Helixer) for structural prediction and splicing assessment.

Main Results:

• Identified 605 confirmed cases of chimeric mis-annotations across the studied genomes.
• The majority of errors were found in invertebrate and plant genomes.
• Machine-learning tools demonstrated efficacy in identifying these mis-annotations.

Conclusions:

• Chimeric mis-annotations are prevalent in genomic datasets, impacting data reliability.
• Machine-learning tools like Helixer offer a promising approach to refine gene models.
• Correcting these errors enhances the understanding of non-model organism genomes.