Additional file 6: Figure S4. of CPAG: software for leveraging pleiotropy in GWAS to reveal similarity between human traits links plasma fatty acids and intestinal inflammation

Relationship between the p value threshold for SNP inclusion and trait clustering. a Similar clusters for obesity (I), autoimmunity (II), and atherosclerosis (III) are observed with the different p value thresholds. Therefore, the detection of informative clusters is robust to varying the number of SNPs in the analysis, although traits are lost as the p value threshold is made more stringent. In all cases, the number of clusters (k) in the analysis was set to 20. The locations of the three clusters were also marked in the entire dendrogram and heat map of pairwise human traits for the p value threshold of 10â 5 (b), 10â 7 (c) and 10â 10 (d). (ZIP 3188 kb)

用于单核苷酸多态性（Single Nucleotide Polymorphism, SNP）纳入筛选的p值阈值与性状聚类之间的关联。图a：在不同p值阈值下，均可观测到肥胖（I）、自身免疫病（II）与动脉粥样硬化（III）的相似聚类结果。因此，尽管随着p值阈值愈发严格，会有部分性状被剔除，但有效聚类的检测结果对于分析中纳入的SNP数量变化具有稳健性。所有分析场景中，聚类数（k）均固定为20。针对p值阈值分别为10⁻⁵（图b）、10⁻⁷（图c）与10⁻¹⁰（图d）的全人类两两性状的完整聚类树状图与热图，也均标注出了上述三类聚类的位置。（ZIP 3188 KB）