Fig 1.
Map of included study sites in the Peruvian Amazon.
A. Provinces in the Loreto region. Three provinces were included in this study (Loreto: orange, Maynas: green, Mariscal Ramon Castilla: blue). Loreto province, includes samples from the Trompeteros district. The remote indigenous Nueva Jerusalen community is settled there. In Mariscal Ramon Castilla province, samples were included from 3 districts containing border communities with Colombia and Brazil. The area within the red square is shown in B. B. Maynas province, with many districts with peri-urban communities. Maps were created with an in-house script in R using the R-packages used R libraries SF, ggspatial, ggrepel. The base layer of the map was obtained from: https://ide.inei.gob.pe/#geo.
Fig 2.
Overview of samples from Peru over time (year) and district.
Samples were collected in 11 districts (y-axis) between 2007 and 2020 (year on x-axis). The size of the bubbles indicates the number of samples collected in a specific year and district.
Fig 3.
Principal Component Analysis of P.
vivax genomes dataset (23), filtered on the SNPs in the Pv Peru Barcode only (A & C), and all biallelic SNPs detected by the Pv AmpliSeq v2 Peru assay in the Peruvian Amazon samples (B&D). The samples (dots) are colored according to the originating population (regions or country), using the classifications from Adam et al., 2022 [68].
Fig 4.
Molecular markers for transmission intensity.
Violin plot of Expected heterozygosity (He) of 40-SNP barcode positions (A & B), and nucleotide diversity (pi) (C and D) measured across the targeted region in 5000 bp windows by district (A& C) and years (B & D).
Fig 5.
Population structure and genetic differentiation.
Heatmap of pairwise FST values between 5 districts in the Peruvian Amazon estimated with the hierfstat package in R (A). Scatter plot of discriminant analysis eigenvalues 1 and 2 (LD1 and LD2) using all variants detected by the Pv AmpliSeq v2 Peru assay of Peru samples (n = 230) grouped by district (B) or district and year (C), and Maynas samples (n = 124) grouped by period (D).
Fig 6.
Network of individual relatedness at (A) intermediate levels of relatedness (50% IBD threshold) and (B) very high levels of relatedness, indicating clonal infections (95% IBD threshold) colored by district. (C) Network of connectivity showing pairwise IBD-sharing between isolates using low levels of relatedness (10% IBD threshold) and summing pairwise comparisons by districts.
Fig 7.
Prediction of origin of P.vivax.
The origin of samples was predicted using the vivaxgen geo framework (https://geo.vivaxgen.org/), which returns 3 predictions of origin for each sample with a score for each prediction. Here we plot a heatmap of the frequency (y-axis) of the highest-scoring prediction of origin for the samples from each district of the sample collection site (x-axis). The majority of samples from Mazan, San Juan Bautista, Iquitos and Trompeteros were predicted to originate from Peru, while bordering communities Ramon Castilla, San Pablo and Yavari were predicted to originate from Brazil.