Figure 1

Algorithmic framework for calculation of association scores. This flowchart outlines the basic algorithm common to the association methods examined. All variations use as input a vector of phenotype values across multiple mouse strains (e.g., HDLC levels, sweet taste preference) and a genotype matrix of allele calls across our 11 K SNP set. At each position in the genome, these algorithms use the local genotype data to generate inferred haplotypes. A test statistic is computed and its significance is estimated to detect groups with different mean phenotypes. This framework is used to evaluate our (A) single-marker method (SMM), (B) parametric inferred haplotype method (IH-P), (C) Kruskal-Wallis based inferred haplotype method (IH-KW), and (D) bootstrap inferred haplotype method (IH-B).