The Protein-Coding Human Genome: Annotating High-Hanging Fruits.

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Human gene annotation relies on manual curation, transcript data, and proteomics, but accuracy is limited for ~2000 genes and alternative transcripts.
• Distinguishing functional expression from noise using transcriptome mapping alone is challenging.
• Previous algorithms addressed micro-exons and wobble splice variants, but limitations remained.

Purpose of the Study:

• To develop and evaluate an algorithm that incorporates splice mechanism and protein structure knowledge for improved exon prediction.
• To re-evaluate gene annotations and identify novel coding exons and splice mechanism re-assignments.
• To address the need for distinctive annotations in the context of the emerging human pan-genome.

Main Methods:

• Developed a novel algorithm integrating splice mechanism and protein structure to predict neighboring homologous exons, including mutually exclusive ones.
• Evaluated predicted exons using transcript data, structural compatibility, and evolutionary conservation.
• Incorporated knowledge of protein structure and splice mechanisms into the prediction pipeline.

Main Results:

• Identified hundreds of novel coding exons, significantly improving gene annotation accuracy.
• Led to re-assignments of splice mechanisms for numerous genes.
• Demonstrated the effectiveness of integrating structural and mechanistic information for exon prediction.

Conclusions:

• The developed algorithm enhances the accuracy of human gene annotation, particularly for complex cases involving alternative splicing.
• Integrating protein structure and splice mechanisms provides a powerful approach to overcome limitations of transcript-based methods.
• Accurate gene annotation is crucial for understanding human genetic diversity within the pan-genome context.