ExTraMapper: Exon- and Transcript-level mappings for orthologous gene pairs

Ferhat Ay, Abhijit Chakraborty, Ramana V. DavuluribioRxiv 2018 · DOI

TL;DR

This paper introduces ExTraMapper, a novel computational method that precisely maps orthologous exons and transcripts between species, going beyond traditional gene-level comparisons. By leveraging sequence conservation and gene annotations, ExTraMapper uncovers a richer landscape of evolutionary relationships, including previously missed microexons and complex splicing events, leading to more accurate transcript-level comparisons and improved downstream analyses.

What Problem Does This Paper Address?

The study of gene evolution has historically focused on gene-level orthology, overlooking the significant diversity generated by alternative splicing and alternative promoter usage within genes. This "one gene, one protein" paradigm fails to capture the full picture of functional evolution, especially in mammals where a single gene can produce hundreds of transcript isoforms. Existing methods for comparing transcripts between species often struggle with this complexity, leading to an incomplete understanding of how gene function evolves at the transcript level. Specifically, the authors highlight the challenge of accurately mapping orthologous exons and transcripts, particularly for smaller exons (microexons) and complex splicing events, which are often missed by current tools. This gap hinders detailed evolutionary studies and the translation of findings from model organisms to humans.

Methods

ExTraMapper employs a multi-step approach:

  1. Exon-level Similarity Computation: It first maps exons between orthologous genes using the liftOver tool to account for genomic coordinate differences. A pairwise exon similarity score is then calculated based on the overlap between lifted-over exon coordinates and their original lengths, considering both overall and coding regions.
  2. Exon Classification: Exons are classified based on their similarity scores and splice site information (acceptor/donor sites) into categories like "conserved," "low similarity," "exon lost," or "exon split/region deleted."
  3. Transcript-level Similarity Computation: Using the computed exon similarities, ExTraMapper calculates overall and coding similarity scores for all possible transcript pairs within an orthologous gene pair. A greedy approach is used, prioritizing highly conserved exon pairings to avoid biases from multiple low-quality mappings.
  4. Tie-breaking and One-to-One Mapping Extraction: To resolve one-to-many or many-to-many transcript relationships, a series of tie-breaking rules are applied, prioritizing coding similarity, then overall similarity, followed by differences in coding exon counts and then overall exon counts. This aims to identify the most likely one-to-one orthologous transcript pairs.

Datasets

The study utilizes publicly available datasets:

  • Gene Annotations and Orthologous Pairs: Ensembl database release 81 for human (Homo_sapiens.GRCh38.81) and mouse (Mus_musculus.GRCm38.81) genomes, along with Ensembl orthology files. This dataset covers approximately 16,711 gene pairs, with 15,846 annotated as protein-coding.
  • Conservation Scores: MAF (Multiple Alignment Format) files from UCSC Genome Browser, processed with maf_parse and PhastCons, to compute conservation scores for human and mouse exons.
  • Exon Mappings (for comparison): Mappings from OrthoExon (Fu et al.) and Inparanoid (Zhang et al.) were downloaded and processed.
  • Transcript Mappings (for comparison): Mappings from Exalign (Pavesi et al.) were used.
  • RNA-seq Data: Human and mouse tissue-specific RNA-seq data from the ENCODE project (FASTQ format) was used for expression profile analysis.

Key Insights and Ideas

  • Beyond Gene-Level Orthology: The paper strongly advocates for and demonstrates the necessity of moving beyond gene-level orthology to understand evolutionary dynamics, especially given the prevalence of alternative splicing.
  • Microexons are Important: ExTraMapper's ability to identify mappings for microexons (exons < 21 bp) is a significant advancement, as these are often missed by other methods and are known to be involved in critical protein functions.
  • Splicing Variation is Evolutionary Information: The method explicitly identifies exon fusions, splits, and losses due to splice site mutations, highlighting how these events contribute to evolutionary divergence.
  • Transcript Length Bias in Existing Methods: The study reveals that methods relying solely on exon lengths (like Exalign) exhibit a strong bias towards longer transcripts, potentially misrepresenting evolutionary relationships.
  • Improved Transcript Expression Correlation: More accurate transcript mappings lead to better correlations in expression profiles between orthologous transcripts in human and mouse tissues, suggesting ExTraMapper provides a more biologically relevant representation.

Key Results

  • Exon Mappings: ExTraMapper identified over 250,000 exon mappings between human and mouse. This is significantly more than previous methods like OrthoExon and Inparanoid, especially for perfect exon mappings (100% conservation).
  • Microexon Mappings: ExTraMapper mapped substantially more microexons (939 human, 829 mouse with <21 bp threshold) compared to OrthoExon (<100) and Inparanoid (<350).
  • Transcript Mappings: The method identified over 30,000 transcript mappings. For example, it reported 30,013 human transcripts mapping to a single mouse transcript with a similarity score > 0.8, compared to 10,405 for Exalign.
  • One-to-One Mappings: ExTraMapper significantly reduces one-to-many mappings, reporting only 177 human transcripts mapping to multiple mouse isoforms, whereas Exalign reported 13,462.
  • Expression Correlation: Transcript pairs identified by ExTraMapper showed higher expression correlation between matched human and mouse tissues compared to those identified by Exalign.
  • BRAF Gene Case Study: The method successfully identified complex exon fusions, splits, and losses in the BRAF gene, providing a more detailed and accurate transcript-level orthology than Ensembl.

Limitations and Future Work

  • Annotation Quality: The number of mouse transcripts is noted as a potential bottleneck for transcript mapping, possibly reflecting differences in biological diversity or annotation quality between human and mouse.
  • Computational Complexity: While not explicitly stated as a limitation, the multi-step process and the need to compare all possible transcript pairs suggest a computationally intensive approach.
  • Future Directions: The authors suggest further research into better annotation of transcript mappings and extending ExTraMapper to include multiple mammalian genomes simultaneously. They also highlight the potential of ExTraMapper as a dictionary for translating transcript-level information between species, particularly for translational research using model organisms.

Summary updated

Abstract

ABSTRACT Access to large-scale genomics and transcriptomics data from various tissues and cell lines allowed the discovery of wide-spread alternative splicing events and alternative promoter usage in mammalians. However, evolutionary studies primarily focus on gene-level orthology relationships, which hinders the importance of transcript-level diversity. Between human and mouse, gene-level orthology is currently present for nearly 16k protein-coding genes spanning a diverse repertoire of over 200k total transcript isoforms. Here we describe a novel method, ExTraMapper, which leverages sequence conservation between exons of a pair of organisms and identifies a fine-scale orthology mapping at the exon and then transcript level. ExTraMapper identifies more than 250k exon, as well as 30k transcript mappings between human and mouse using only sequence and gene annotation information. We demonstrate that ExTraMapper identifies a larger number of exon and transcript mappings compared to previous methods. Further, it identifies exon fusions, splits, and losses due to splice site mutations, and finds mappings between microexons that are previously missed. By reanalysis of RNA-seq data from 13 matched human and mouse tissues, we show that ExTraMapper improves the correlation of transcript-specific expression levels suggesting a more accurate mapping of human and mouse transcripts. ExTraMapper also reports better transcript-level mappings compared to Ensembl orthology for the human proto-oncogene BRAF and its mouse ortholog as well as several other example genes with important isoform-specific functions. ExTraMapper is applicable to any pair of organisms that have orthologous gene pairs and is available at https://github.com/ay-lab/ExTraMapper and http://ay-lab-tools.lji.org/extramapper
BibTeX
@misc{ay2018extramapper,
  author = {Ferhat Ay and Abhijit Chakraborty and Ramana V. Davuluri},
  title = {{ExTraMapper: Exon- and Transcript-level mappings for orthologous gene pairs}},
  year = {2018},
  doi = {10.1101/277723},
  howpublished = {bioRxiv},
}

Loading related content...

Figure