This definition of the weights of the group graph expressed between subject cluster-similarity in such a way that high weight values reflected strong subject overlap.
Figure S2 shows the HCV NS3 and NS5B phylogenetic trees with the well-supported individual subject clusters and separation into genotype 1A and 1B groups.
To identify influential subject clusters, we tested random slopes models as well as individual autoregressive models.
Within each cell type, cell lines derived from the same subject clustered more closely together than to the cell lines from different subjects.
We conducted residual diagnostics to assess deviations from assumed functional form, and the presence of influential data points and subject clusters.
We conducted residual diagnostics to assess the presence of influential data points and subject clusters, as well as deviations from assumed functional form.
Random slopes and individual autoregressive models showed small, highly influential subject clusters (seven subjects) with positive associations between air pollutants and antioxidant enzymes, whereas most of the remaining 53 subjects showed inverse associations.
Of note is that healthy control subjects cluster together at the left side of the graph.
The PCA also gives the opportunity to multivariately investigate how subjects cluster into subgroups and how certain subgroups differ.
This approach groups masses based on their similarity of intensity between samples, meaning that masses with related fold-changes between subjects cluster together, independent of their absolute levels.
The 48 African American subjects cluster with HapMap African American samples on the principal components analysis (PCA) plot, indicating mixed African and European ancestry.
