Figure 2

The artificial imputation implemented in SnipSnip increased the association signal compared with single-SNP analysis.
Manhattan plots show P values of SNPs (y-axis, -log10 scale) on a genomic scale (x axis) of A(H7N9) GWAS dataset. The single-SNP allelic association P values using logistic regression implemented in PLINK and artificial imputation P values using SnipSnip are shown in the upper and lower panel respectively.