Figure 1

Site frequency spectra of demographic models and data with a sample size of 900. The SFS for 3 demographic models, the Neutral Regions (NR) data and 7 categories of the Exome Sequencing Project (ESP) data. To adjust for the different sample sizes in the two datasets, probabilistic subsampling was applied to make all sample sizes equal to 900 chromosomes. Only the first 10 minor allele count categories are shown. For each minor allele count, from left to right: constant population size, European history with 2 bottlenecks but no growth [10], European history with recent growth (Model II in [3]), the NR data, intergenic SNVs of the ESP data, intron SNVs of the ESP data, synonymous SNVs of the ESP data, UTR SNVs of the ESP data, missense SNVs of the ESP data, nonsense SNVs of the ESP data and splice SNVs of the ESP data.