Data characteristics for the 1000 genomes exon sequencing pilot datasets. A. Principal component analysis of a “mixed” read depth matrix built with data from 3 different sequencing centers, SC (Wellcome Trust Sanger Institute), BI (Broad Institute) and BCM (Baylor College of Medicine). Each sample is represented as a point in the plot, with the first principal component plotted vs. the second principal component. Samples from different sequencing centers cluster separately from each other within this space, suggesting significant differences in the gene affinities among these three datasets. B. Distributions of MRD for each of the BCM, BI and SC samples C. Histogram of RD across all GSSs in the three datasets. D. Histogram of gene affinities across genes within each of the three datasets. E. Distributions of the RD over-dispersion factor (ODF) in our data.