Table 1.
Gene regions from which SNPs were extracted.
Table 2.
Numbers of SNPs identified in VEEV genomes, by gene region.
Fig 1.
SNP phylogeny of VEEV isolates by parsimony.
Strains are labeled by serotype-country-year collected-strain-host. Country codes are GA = Guatemala, PE = Peru, NI = Nicaragua, VE = Venezuela, CO = Colombia, TR = Trinidad, PA = Panama, US = USA, EC = Ecuador, ME = Mexico, BE = Belize, HO = Honduras, BR = Brazil, AR = Argentina, FG = French Guiana. Host codes are hor = horse, don = donkey, hum = human, mos = mosquito, ham = hamster, mus = mouse. u = unknown. Strains are colored by serotype (blue = IE, green = ID, red = IC, and purple = IAB). Hosts from which the strains were collected are indicated with symbols at the branch tips (red circles = human, orange circles = horses, blue circles = mosquitos, and green squares = hamsters). Counts of the number of alleles shared uniquely by the sequences down each branch are shown at the nodes in blue.
Fig 2.
Decision tree for prediction of serotype from SNP alleles.
Notations above internal nodes indicate SNP position in the TC-83 genome and alleles corresponding to left and right branches. Numbers below terminal nodes are numbers of isolates in node with serotypes IAB/IC/ID/IE respectively.
Fig 3.
Decision tree for prediction of host type from SNP alleles.
Notations above internal nodes are as in Fig 2. Numbers below terminal nodes are numbers of isolates in node collected from large/small host types, respectively.
Table 3.
Accuracy, positive (PPV) and negative predictive value (NPV), true positive (TPR) and negative (TNR) rates for serotype and host type predictions.
Table 4.
Comparison of trees from multiple sequence alignment versus all SNPs, and trees from SNPs located in a single gene versus all SNPs.