HBA-DEALS: accurate and simultaneous identification of differential expression and splicing using hierarchical Bayesian analysis

Guy Karlebach,Sheng Li,Peter N Robinson,Peter N Robinson,Robin Steinhaus,Robin Steinhaus,Olga Anczukow,Diogo Ft Veiga,Peter Hansen,Daniel Danis,Olga Anczukow

doi:10.1186/s13059-020-02072-6

Abstract

We present Hierarchical Bayesian Analysis of Differential Expression and ALternative Splicing (HBA-DEALS), which simultaneously characterizes differential expression and splicing in cohorts. HBA-DEALS attains state of the art or better performance for both expression and splicing and allows genes to be characterized as having differential gene expression, differential alternative splicing, both, or neither. HBA-DEALS analysis of GTEx data demonstrated sets of genes that show predominant DGE or DAST across multiple tissue types. These sets have pervasive differences with respect to gene structure, function, membership in protein complexes, and promoter architecture.

Highlights

RNA sequencing (RNA-seq) has become the most commonly used genomic technique for the transcriptome-wide analysis of differential expression and alternative splicing of mRNAs
We show with our method, using data from the Genotype-Tissue Expression (GTEx) project [15], that genes can be assigned to four groups according to the propensity of a gene to display Differential gene expression (DGE), DAST, both, or neither in comparisons between different tissue types
Here, we present a method for joint modeling of differential expression and splicing, Hierarchical Bayesian Analysis of Differential Expression and ALternative Splicing (HBA-DEALS)

Summary

Introduction

RNA sequencing (RNA-seq) has become the most commonly used genomic technique for the transcriptome-wide analysis of differential expression and alternative splicing of mRNAs. Since its introduction over a decade ago, Illumina short-read sequencing technology has been the dominant platform for carrying out RNA-seq experiments, but newer long-read single-molecule sequencing technologies of Pacific Biosciences and Oxford Nanopore provide alternatives that may allow a more accurate and comprehensive assessment of isoform diversity [1]. Analysis of RNA-seq data is done in a pipeline that maps raw reads to genes or isoforms (transcripts), quantifies the number (count) of reads associated with each isoform, generating an expression matrix, followed by normalization steps and statistical analysis of differential expression [2, 3]. Differential gene expression (DGE) refers to alterations in the expression (counts) of the sum of each of the isoforms that are encoded by a gene. Many methods for DGE analysis are based on discrete probability distributions such as Karlebach et al Genome Biology (2020) 21:171 the Poisson or negative binomial [4, 5]. voom instead estimates the mean-variance relation non-parametrically from log-counts per million reads, which are used as input for linear modeling and empirical Bayes differential expression analysis [6]

Methods

Results

Discussion

Conclusion