Figure 4.

Histograms showing the distribution of all N_inf×n direction cosine values. {α_ij:1≤i≤N_inf,1≤j≤n} for different multi-allelic data sets. Only samples judged to be infected based on the magnitude histogram (see Figure 1) are included; the number of infected samples is denoted N_inf, and the number of alleles for a given data set is denoted n. The dashed vertical line indicates the diagnosis threshold for infection vs. noninfection of individual alleles. For the multi-allellic data sets, the thresholds were determined by visual inspection of the histograms. The pvdhfr 57-58-61 data contained eight alleles, giving 338 × 8 = 2704 direction cosines total in the histogram; direction cosine threshold = 0.2. Histogram bin width = 0.02. Inset panel shows magnified region near the diagnostic threshold, including small vertical blue tick marks indicating the direction cosines for individual alleles. Corresponding figures for pvdhps 382-383 (with four alleles) and for pvdhfr 117 (with three alleles) are available as Figures S2 and S4, respectively (See Additional Files 4 and 5).