Heritability of alternative splicing in the human genome

Abstract

Alternative pre-mRNA splicing increases proteomic diversity and provides a potential mechanism underlying both phenotypic diversity and susceptibility to genetic disorders in human populations. To investigate the variation in splicing among humans on a genome-wide scale, we use a comprehensive exon-targeted microarray to examine alternative splicing in lymphoblastoid cell lines (LCLs) derived from the CEPH HapMap population. We show the identification of transcripts containing sequence verified exon skipping, intron retention, and cryptic splice site usage that are specific between individuals. A number of novel alternative splicing events with no previous annotations in either the RefSeq and EST databases were identified, indicating that we are able to discover de novo splicing events. Using family-based linkage analysis, we demonstrate Mendelian inheritance and segregation of specific splice isoforms with regulatory haplotypes for three genes: OAS1, CAST, and CRTAP. Allelic association was further used to identify individual SNPs or regulatory haplotype blocks linked to the alternative splicing event, taking advantage of the high-resolution genotype information from the CEPH HapMap population. In one candidate, we identified a regulatory polymorphism that disrupts a 5' splice site of an exon in the CAST gene, resulting in its exclusion in the mutant allele. This report illustrates that our approach can detect both annotated and novel alternatively spliced variants, and that such variation among individuals is heritable and genetically controlled.

Figure 1.

(A) Schematic for coverage of probe sets across the entire length of the transcript. Yellow regions are exons, whereas gray regions represent introns. The short dashes underneath the exon regions indicate individual probes of 25 nucleotides in length representing the probe set. The Affymetrix GeneChip Human Exon 1.0 ST Array allows for exon-level expression profiling in a single chip, and can interrogate over one million predicted exons within the human genome. (B) Flowchart for processing and analysis of chips to validation of alternative splicing events. Total RNA is extracted from the two cell lines (n = 15 replicates per individual) and is transcribed to cDNA and labeled with biotin. The total cDNA is then hybridized to the exon chip, followed by washing and staining with an anti-streptavidin antibody. Chips are then scanned, and hybridization data are processed and analyzed by the Affymetrix Power Tools (version 1.6) software package. A splicing index is calculated for ∼1.4 million probe sets covering one million exons. A subset of 20 alternative splicing events predicted between the two individuals using an unpaired t-test (P < 8.915 × 10⁻⁴) on the splicing index and other criteria (see Methods), are then validated by (1) RT-PCR using exon body primers flanking the probe set of interest and (2) sequencing of the RT-PCR products.

Figure 2.

Principal component analysis. A three-dimensional plot of the splicing index data showing the three passages of five technical replicates each of individuals GM12750 and GM12751, on the left and right sides, respectively. The three biological replicates are shown as purple, orange, and yellow spheres, respectively. The three outliers that were removed from all subsequent analyses are shaded in a blue sphere. The percentage of variance attributed to principal components one and two is shown on the X- and Y-axes, respectively. Plots were created using the Partek Genomics Suite software package (Partek).

Figure 3.

Heritability of alternative splicing. Inheritance of alternative splicing for genes (A) OAS1, (B) CRTAP, and (C) CAST. Left panel shows pedigree structure of CEPH/UTAH family 1444 with the autosomal dominant inherited splice pattern as blue symbols. Haplotypes for each of the eight founder chromosomes are labeled A, B, C, D, E, F, G, and H, and the two inherited haplotypes of each family member are indicated within the symbol. The regulatory haplotype is shown as bold white text. Squares represent males, and circles represent females. CEPH/UTAH 1444 pedigree is labeled as follows: 1 (GM12739), 2 (GM12740), 3 (GM12741), 4 (GM12742), 5 (GM12743), 6 (GM12744), 7 (GM12745), 8 (GM12746), 9 (GM12847), 10 (GM12747), 11 (GM12748), 12 (GM12749), 13 (GM12750), and 14 (GM12751). The right panel shows the two transcript isoforms of the genes. Exon-body primers are shown above the flanking exons of the predicted alternatively spliced exons. Shown below the transcript isoforms are the RT-PCR results. Lanes are numbered from 1–14 according to the pedigree on the left.

Figure 4.

Association of alternative splicing and genotypes for the CAST gene. (A) RT-PCR of CAST exon against a panel of unrelated parents from each of the 30 HapMap CEU trios. Sample names are colored according to their genotype for SNP rs7724759: homozygous GG (green), homozygous AA (red), and heterozygous AG (black). (B) Four known isoforms of the CAST gene are shown with their RefSeq accession numbers on the left and the candidate probe set shaded in gray. Shown below is the sequence of the exon in capital letters and flanked by the intronic sequence in lower case. The SNP rs7724759 is located at the last position of the exon and is a G to A substitution that disrupts the consensus splice site sequence.