Dynamic Scan Procedure for Detecting Rare-Variant Association Regions in Whole Genome Sequencing Studies

Abstract

1Whole genome sequencing (WGS) studies are being widely conducted to identify rare 2 variants associated with human diseases and disease-related traits. Classical single-3 marker association analyses for rare variants have limited power, and variant-set based 4analyses are commonly used to analyze rare variants. However, existing variant-set 5 based approaches need to pre-specify genetic regions for analysis, and hence are not 6 directly applicable to WGS data due to the large number of intergenic and intron regions 7 that consist of a massive number of non-coding variants. The commonly used sliding 8 window method requires pre-specifying fixed window sizes, which are often unknown as 9 a priori, are difficult to specify in practice and are subject to limitations given genetic 10 association region sizes are likely to vary across the genome and phenotypes. We 11 propose a computationally-efficient and dynamic scan statistic method (Scan the 12 Genome (SCANG)) for analyzing WGS data that flexibly detects the sizes and the 13 locations of rare-variants association regions without the need of specifying a prior fixed 14 window size. The proposed method controls the genome-wise type I error rate and 15 accounts for the linkage disequilibrium among genetic variants. It allows the detected 16 rare variants association region sizes to vary across the genome. Through extensive 17 simulated studies that consider a wide variety of scenarios, we show that SCANG 18 substantially outperforms several alternative rare-variant association detection methods 19 while controlling for the genome-wise type I error rates. We illustrate SCANG by 20 analyzing the WGS lipids data from the Atherosclerosis Risk in Communities (ARIC) 21 study. 22 23

PSCAN: Spatial scan tests guided by protein structures improve complex disease gene discovery and signal variant detection

“…In the setting of scan methods in 1D, this signal detection approach can achieve asymptotic optimality [ 24 , 47 ] (i.e., in reliably separating the true signal region from noise) when signals are sufficiently strong and signal regions are well separated. However, an alternative signal region identification approach has been developed to deal with situation where signals are relatively weak and/or signal regions are possibly nested [ 48 , 49 ]. This procedure only removes windows that overlap by more than the pre-specified overlap fraction f .…”

eSCAN: Scan Regulatory Regions for Aggregate Association Testing using Whole Genome Sequencing Data

“…The eSCAN procedure can be split into two steps: a p-value computing step and a decision-making step (using a p-value threshold). First, for each enhancer, set-based p-values are calculated by fastSKAT, which applies randomized singular value decomposition (SVD) to rapidly analyze much larger regions than standard SKAT, and then p-values are "averaged" by the Cauchy method via ACAT 4 . Second, eSCAN calculates two types of significance threshold.…”

“…While we do use a classic Bonferroni correction in our real data example from WHI, due to the small sample size available to us for replication, this is almost certainly overconservative. eSCAN provides two estimations of significance threshold, either empirically or analytically, using the strategies from SCANG and WGScan respectively, which have demonstrated significant enrichments of signals in Li et al4 and He et al10 . In addition, although our analyses focused on unrelated individuals, it can be readily extended to related samples by replacing the generalized linear model (GLM) with the generalized linear mixed model (GLMM) in the first step 4 .One potential limitation of eSCAN is the lack of base pair resolution in defining regions important for gene regulation, due to the sparsity of reads with most Hi-C and chromatin conformation assays (leading to resolution as broad as 40 kb when assessing interactions between genomic regions).…”

Identifying rare variants for quantitative traits in extreme samples of population via Kullback-Leibler distance

“…Thus costeffective design should be considered to reduce sample size. Another challenge is that the statistical power with test statistics of single-marker tests is generally low in genetic association studies of rare variants with more moderate or weak genetic effects [8][9][10]. To date many statistical methods have been developed for rare variant association analysis, including burden tests [11][12][13], variance-component tests [14,15], series of sequence kernel association tests [10,16,17].…”

“…To date many statistical methods have been developed for rare variant association analysis, including burden tests [11][12][13], variance-component tests [14,15], series of sequence kernel association tests [10,16,17]. Any of these methods has relative perfect performance in special scenario, but none of them can overwhelm others in all scenarios [8,9], especially for quantitative traits.…”