Predicting functional alternative splicing by measuring RNA selection pressure from multigenome alignments

Abstract

High-throughput methods such as EST sequencing, microarrays and deep sequencing have identified large numbers of alternative splicing (AS) events, but studies have shown that only a subset of these may be functional. Here we report a sensitive bioinformatics approach that identifies exons with evidence of a strong RNA selection pressure ratio (RSPR)--i.e., evolutionary selection against mutations that change only the mRNA sequence while leaving the protein sequence unchanged--measured across an entire evolutionary family, which greatly amplifies its predictive power. Using the UCSC 28 vertebrate genome alignment, this approach correctly predicted half to three-quarters of AS exons that are known binding targets of the NOVA splicing regulatory factor, and predicted 345 strongly selected alternative splicing events in human, and 262 in mouse. These predictions were strongly validated by several experimental criteria of functional AS such as independent detection of the same AS event in other species, reading frame-preservation, and experimental evidence of tissue-specific regulation: 75% (15/20) of a sample of high-RSPR exons displayed tissue specific regulation in a panel of ten tissues, vs. only 20% (4/20) among a sample of low-RSPR exons. These data suggest that RSPR can identify exons with functionally important splicing regulation, and provides biologists with a dataset of over 600 such exons. We present several case studies, including both well-studied examples (GRIN1) and novel examples (EXOC7). These data also show that RSPR strongly outperforms other approaches such as standard sequence conservation (which fails to distinguish amino acid selection pressure from RNA selection pressure), or pairwise genome comparison (which lacks adequate statistical power for predicting individual exons).

Figure 1. Prediction of functional alternative exons by RNA Selection Pressure Ratio.

(A) The alternative splicing pattern of GRIN1 in human; (B) RNA selection pressure ratio (RSPR) of each exon (measured vs. all conserved exons as the control set); (C) Synonymous mutation rates Ks for the constitutive exons (in blue) and the alternative exons (in red) on each branch of the phylogenetic tree. To fit the available space, only 7 of the 16 species in the computed tree are shown; the other species show the same pattern; (D) RESCUE ESE and FAS ESS predictions for the C1 and N1 exons.

Figure 2. Sensitivity of detection of known NOVA splicing regulator targets by RSPR.

Sensitivity of detection of known NOVA splicing regulator targets by five methods: (A) RSPR calculated for each alternative exon vs. constitutive exons as a control; (B) RSPR calculated for each alternative exon vs. all other exons in the gene; (C) RSPR calculated using just the human vs. mouse genomes; (D) mutation ratio calculated using baseml from the complete set of vertebrate genomes (see text); (E) sequence conservation calculated using phastCons from the complete set of vertebrate genomes (see text).

Figure 3. Validation of RSPR by independent alternative splicing data from other species.

(A) The fraction of exons that were observed to be alternatively spliced in independent EST data for different species (y-axis), as a function of RSPR (x-axis). (B) The prediction performance (true positive rate vs. upper bound estimate of false positive rate) for RSPR vs. sequence conservation calculated using baseml or phastCons.

Figure 4. Comparing the predictive value of RSPR with standard sequence conservation.

(A) The fraction of exons that were observed to be alternatively spliced in independent mouse EST data (y-axis), as a function of RSPR (x-axis), under different mutation ratio constraints (see text) (B) The fraction of exons that were observed to be alternatively spliced in independent mouse EST data (y-axis), as a function of baseml exon mutation ratio (x-axis), under different RSPR constraints (RSPR>3, high RNA selection pressure; RSPR<3, low RNA selection pressure). (C) The baseml intron mutation ratio (based on standard sequence conservation in the adjacent 50 nt of intron flanking the alternative exon; y-axis), as a function of RSPR (x-axis). Error bars represent one standard deviation.

Figure 5. Experimental validation of RSPR predictions.

Tissue-specific splice regulation is one indicator of functional alternative splicing that can easily be assayed experimentally. (A) Summary of RT-PCR results for 20 high-RSPR exons and 20 low-RSPR exons (see text). (B) RT-PCR detection of GIT1 splicing in 10 human tissues shows that this alternative exon is regulated in a brain-specific manner. (C) Example of RT-PCR detection of non-tissue-specific AS. (D) Example of RT-PCR detection of constitutive splicing.

Figure 6. RSPR analysis of alternative splicing inclusion levels.

(A) The distribution of the RNA selection pressure ratio (RSPR) for different inclusion levels (see text for detailed explanation); (B) The fraction of exons that were observed to be alternatively spliced in mouse EST data (y-axis) vs. RSPR; (C) The fraction of exons that preserve the protein reading frame, as a function of RSPR. Error bars represent one standard deviation.

Figure 7. Prediction of a novel functional AS event in EXOC7.

(A) RSPR detects strong selection in one AS exon (red), conserved in human, mouse, dog, cow and frog. (B) The exon encodes 13 aa in a loop region between helix 12 and helix 13, adjacent to a set of predicted phosphorylation sites. (C) Exonic Splicing Silencer (ESS) motifs predicted by FAS-ESS within the exon. (D) RT-PCR analysis of EXOC7 splicing in 10 human tissues. (E) Conservation of ESS sequences (black) within the exon alignment.