Wavelet-based genetic association analysis of functional phenotypes arising from high-throughput sequencing assays

Abstract

Understanding how genetic variants influence cellular-level processes is an important step toward understanding how they influence important organismal-level traits, or “phenotypes,” including human disease susceptibility. To this end, scientists are undertaking large-scale genetic association studies that aim to identify genetic variants associated with molecular and cellular phenotypes, such as gene expression, transcription factor binding, or chromatin accessibility. These studies use high-throughput sequencing assays (e.g., RNA-seq, ChIP-seq, DNase-seq) to obtain high-resolution data on how the traits vary along the genome in each sample. However, typical association analyses fail to exploit these high-resolution measurements, instead aggregating the data at coarser resolutions, such as genes, or windows of fixed length. Here we develop and apply statistical methods that better exploit the high-resolution data. The key idea is to treat the sequence data as measuring an underlying “function” that varies along the genome, and then, building on wavelet-based methods for functional data analysis, test for association between genetic variants and the underlying function. Applying these methods to identify genetic variants associated with chromatin accessibility (dsQTLs), we find that they identify substantially more associations than a simpler window-based analysis, and in total we identify 772 novel dsQTLs not identified by the original analysis.

Wavelet Screening: a novel approach to analysing GWAS data

“…Note that, in contrast to Shim and Stephens (2015), our test statistic applies equal weight to each scale. In Shim and Stephens (2015), lower scales have more weight.…”

“…Note that, in contrast to Shim and Stephens (2015), our test statistic applies equal weight to each scale. In Shim and Stephens (2015), lower scales have more weight. As shown in the upper panel in Figure 2, the separation between the null and the alternative achieved by this test Second, we extract a vertical summary by considering sub-regions of the overall screened region.…”

“…To address these shortcomings, we developed "Wavelet Screening" as a new approach to GWAS by harnessing insights from functional modeling (Morris and Carroll, 2006;Shim and Stephens, 2015) . Specifically, we adapt the approach described by Shim and Stephens (2015) in which the authors test for association with a functional phenotype (the response signal) by first transforming the signal using fast discrete wavelet transform (Mallat, 2008) and then testing for single-SNP associations. In essence, our main idea is to reverse this approach, by modeling the SNP signals as wavelets over a large genomic region and regressing the wavelet coefficients on the phenotype.…”

Simultaneous detection and estimation of trait associations with genomic phenotypes

“…In contrast, the BSmooth method of Hansen and others (2012) and the WaveQTL method of Shim and Stephens (2015) are two-step procedures that leverage the spatial structure of the genomic phenotypes. BSmooth first smooths the data, and then uses the smoothed data to calculate a t-statistic at each site.…”

msCentipede: Modeling heterogeneity across genomic sites improves accuracy in the inference of transcription factor binding

“…Motivated by recent work applying multi-scale methods to analyses of high-throughput sequencing data (Shim and Stephens, 2013; Shim et al, 2014), we use hierarchical multi-scale models to better model heterogeneity in the read profiles across genomic locations and across replicate measurements of chromatin accessibility. Modeling the data at multiple scales explicitly allows us to infer different amounts of genomewide variation at each scale, and enables the automatic identification of relevant scales during inference.…”