Variant Association Tools for Quality Control and Analysis of Large-Scale Sequence and Genotyping Array Data  Gao T. Wang, Bo Peng, Suzanne M. Leal  The American Journal of Human Genetics  Volume 94, Issue 5, Pages 770-783 (May 2014) DOI: 10.1016/j.ajhg.2014.04.004 Copyright © 2014 The American Society of Human Genetics Terms and Conditions

Figure 1 Variant Association Tools Pipeline for Quality Control and Association Analysis. The American Journal of Human Genetics 2014 94, 770-783DOI: (10.1016/j.ajhg.2014.04.004) Copyright © 2014 The American Society of Human Genetics Terms and Conditions

Figure 2 Best Practices for Implementing the Variant Association Tools Pipeline. The American Journal of Human Genetics 2014 94, 770-783DOI: (10.1016/j.ajhg.2014.04.004) Copyright © 2014 The American Society of Human Genetics Terms and Conditions

Figure 3 Global Ancestry Inference Ancestry inference was performed using MDS analysis with graphic presentation of projection of samples to the first two MDS components. Color codes represent the region of collection or self-reported ancestry. MDS analysis was performed for all European and Asian samples (A); European samples (B); and Asian samples (C). The American Journal of Human Genetics 2014 94, 770-783DOI: (10.1016/j.ajhg.2014.04.004) Copyright © 2014 The American Society of Human Genetics Terms and Conditions

Figure 4 Variant- and Sample-Level Summary Statistics for European and Asian Samples by Variant Functional Type: Missense, Synonymous, and Nonsense Variant-level statistics are displayed as the number and percentage of variant sites by functional type (A) and the frequency of variant sites by functional type (B) for Europeans and Asians combined. Sample-level statistics are displayed as average number of variants per individual by functional type for Europeans, Asians, and the two populations combined (C) and the average number of singletons per individual by functional type for European and Asians separately (D). (C) Error bars represent standard error of the mean. The American Journal of Human Genetics 2014 94, 770-783DOI: (10.1016/j.ajhg.2014.04.004) Copyright © 2014 The American Society of Human Genetics Terms and Conditions

Figure 5 Distribution of BMI and Rare-Variant Association Analysis via the CMC Method (A) The simulated BMI values for Europeans (left), BMI values after log10 transformation (center), and quantile normal transformation (right). Analysis of whole-exome association was performed for the quantile-normal-transformed BMI phenotype via the CMC method. For panels (B) to (D), results are represented by p values at the −log10 scale and displayed in quantile-quantile (QQ) (left) and Manhattan (right) plots. Analysis of Europeans (B) and Asians (C) and meta-analysis of European and Asian results (D) was performed via the sample-size-based method. The American Journal of Human Genetics 2014 94, 770-783DOI: (10.1016/j.ajhg.2014.04.004) Copyright © 2014 The American Society of Human Genetics Terms and Conditions