Ranked choice voting for representative transcripts with TRaCE.

Area of Science:

• Genomics
• Transcriptomics
• Bioinformatics

Background:

• Genome sequencing projects generate extensive transcript evidence, but downstream analyses often rely on a single representative transcript per gene locus.
• Selecting a canonical transcript is crucial for consistent and accurate downstream genomic analyses.

Purpose of the Study:

• To introduce Transcript Ranking and Canonical Election (TRaCE), a novel computational method for selecting and ranking gene transcripts.
• To provide a robust framework for identifying canonical transcripts based on comprehensive evidence.

Main Methods:

• TRaCE employs an 'election' system where RNA-seq samples rank transcripts based on annotation edit distance.
• Transcripts are further evaluated using criteria like protein length and InterPro domain coverage.
• A multi-round voting process determines the canonical transcript and orders remaining transcripts by relevance.

Main Results:

• TRaCE successfully identifies canonical transcripts by integrating multiple evidence types.
• The method can prioritize common isoforms from expression atlases or specific alternative transcripts from targeted datasets.
• The election process provides a ranked list of transcripts, offering more nuanced information than a single canonical choice.

Conclusions:

• TRaCE offers a flexible and evidence-based approach to canonical transcript selection.
• The method enhances the accuracy and relevance of transcript data used in downstream genomic studies.
• TRaCE facilitates the analysis of transcriptomes in both broad and specific biological contexts.