Phylogeographic Analysis of Tick-borne Pathogen - Recent Advances and Issues in Environmental Science - page 80

an optimum degree of ecoregion clustering collectively explaining the association of
A. marginale strains with the environment is reached. Such association may be low or
high (in the range 0-1) but is the optimal according to the ecoregion features. After the
use of these methods we retained a total of four ecoregion clusters (Figure 1) to which
A. marginale sequences were unambiguously associated, and the rest were discarded
because A. marginale strains were not isolated there.
Figure 1. World ecoregions. (A) Clusters of vegetation features of the world computed through an
unsupervised classification on Principal Components Analysis decomposition of monthly NDVI
values at a resolution of 0.1°. Small islands were removed to improve the presentation. Only the
ecoregions providing most information about A. marginale distribution according to a Discriminant
Analysis were included, grouped in four main clusters and arbitrarily numbered. Overlying records
of A. marginale are in black. (B) Monthly NDVI values detected for each ecoregion clusters (plotted
with the same colors as in A). (C) Monthly accumulated temperature and rainfall recorded for each
ecoregion cluster.
Analysis of A. marginale MSP1a Repeat Sequences
The conserved amino acid sequences of MSP1a R1 and RL repeats reported in A.
marginale strains from each ecoregion cluster were used to determine the consen-
sus sequence in the cluster. The distances between groups of sequences ascribed to
an ecoregion cluster were then computed, both within and between clusters. The
hypothesis herein is that highest affinities between R1 and RL isolates are observed
when grouped according to the ecoregion to which they are related. An alternative hy-
pothesis is that isolates cluster according to a geographical background. To check for
that alternative hypothesis we examined genetic distances with strains arranged into
geographical groups. We used only American strains to overcome the undersampling
in Africa. They were grouped as Western USA, Eastern USA, Mesoamerica and the