Supplementary Figure S2 from Combination of an Autoantibody Panel and Alpha-Fetoprotein for Early Detection of Hepatitis B Virus-Associated Hepatocellular Carcinoma

Supplementary Figure S2: Samples (n=192) from GSE22058 were clustered and analyzed to detect the presence of outliers. Sample clustering demonstrated that two outliers (GSM548410 and GSM548460) need to be removed and the remaining samples (n=190) were applied for subsequent analyses (Supplementary Figure S2A). The scale-free modules were constructed by the optimal threshold of β=4 (Supplementary Figure S2B). Genes were divided into 16 modules by the thresholding power (Supplementary Figure S2C). The blue module had the highest correlation with the sample traits (r=0.93, P<0.001) (Supplementary Figure S2D) and the genes in this module were highly correlated with clinical information (cor=0.98, P<0.001) (Supplementary Figure S2E).