Figure 6

Strongly connected peptides are concordant when regressed against time. An illustration from the influenza dataset is shown. A protein sub-network consists of peptides mapping to a given protein, forming a all-to-all network with weighted edges. Applying an edge threshold decreases the number of “strong” peptide connections. An edge is counted as concordant if the two connected peptides have significant slopes in the same direction after linear regression against infection day. The discordant edges are clearly differentiated by edge thresholding.