While moderately strong aggregate SNP effects were observed in the full sample by both GCTA and twin biometry, the genome-wide scoring procedure was unable to tap more than a small fraction of that variance. The results from the scoring procedure are given in Figure 3. The top row of sub-figures in Figure 3 provide the genome-wide scoring results for the five phenotypes, under seven SNP proportion thresholds and three LD thresholds. To explain, consider the top right figure. Here we imposed an LD threshold of .05. That is, SNPs were excluded whenever they were in LD > .05 with a nearby, more significant SNP. Each phenotype was then analyzed under seven thresholds for the proportion of SNPs to retain in the score. Stringent thresholds, such as including only .0001 of all SNPs produced essentially null results for every phenotype but nicotine. As the threshold was relaxed and more SNPs were included, improvement is seen for every phenotype (sans nicotine), until the effect plateaus at around .05. This pattern of results is generally true for each of the LD cutoffs (each
