Comparing newly developed SNP barcode panels with microsatellites to explore population genetics of malaria parasites in the Peruvian Amazon

Abstract

Introduction: Malaria molecular surveillance (MMS) can provide insights into transmission dynamics, guiding national control programs. We previously designed AmpliSeq assays for MMS, which include different traits of interest (resistance markers and pfhrp2/3 deletions), and SNP barcodes to provide population genetics estimates of Plasmodium vivax and Plasmodium falciparum parasites in the Peruvian Amazon. The present study compares the genetic resolution of the barcodes in the AmpliSeq assays with widely used microsatellite (MS) panels to investigate population genetics of Amazonian malaria parasites.

Methods: We analyzed 51 P. vivax and 80 P. falciparum samples from three distinct areas in the Loreto region of the Peruvian Amazon: Nueva Jerusalén (NJ), Mazan (MZ), and Santa Emilia (SE). Population genetics estimates and costs were compared using the SNP barcodes (P. vivax: 40 SNPs and P. falciparum: 28 SNPs) and MS panels (P. vivax: 16 MS and P. falciparum: 7 MS).

Results: The P. vivax genetic diversity (expected heterozygosity, He) trends were similar for both markers: He _MS = 0.68-0.78 (p > 0.05) and He _SNP = 0.36-0.38 (p > 0.05). P. vivax pairwise genetic differentiation (fixation index, F_ST) was also comparable: F_ST-MS = 0.04-0.14 and F_ST-SNP = 0.03-0.12 (pairwise p > 0.05). In addition, P. falciparum genetic diversity trends (He _MS = 0-0.48, p < 0.05; He _SNP = 0-0.09, p < 0.05) and pairwise F_ST comparisons (F_ST-MS = 0.14-0.65, F_ST-SNP = 0.19-0.61, pairwise p > 0.05) were concordant between both panels. For P. vivax, no geographic clustering was observed with any panel, whereas for P. falciparum, similar population structure clustering was observed with both markers, assigning most parasites from NJ to a distinct subpopulation from MZ and SE. We found significant differences in detecting polyclonal infections: for P. vivax, MS identified a higher proportion of polyclonal infections than SNP (69% vs. 33%, p = 3.3 × 10^-5), while for P. falciparum, SNP and MS detected similar rates (46% vs. 31%, p = 0.21). The AmpliSeq assay had a higher estimated per-sample cost compared to MS ($183 vs. $27-49).

Discussion: The SNP barcodes in the AmpliSeq assays offered comparable results to MS for investigating population genetics in P. vivax and P. falciparum populations, despite some discrepancies in determining polyclonality. Given both panels have their respective advantages and limitations, the choice between both should be guided by research objectives, costs, and resource availability.

Keywords: AmpliSeq; NGS; SNP; malaria molecular surveillance; microsatellites; population genetics; population structure.

FIGURE 1

Expected heterozygosity (He) in P. vivax and P. falciparum populations using SNP and MS. He was calculated for P. vivax (A, B) and P. falciparum (C, D) using MS (A, C) or SNP (B, D). The horizontal line within the boxes represents the median He from each panel (MS) or position (SNP) for all samples in each group. Dots outside the boxes represent the outlier data. In addition, Spearman’s correlations of He values between MS and the SNP barcode for (E) P. vivax and (F) P. falciparum populations were determined.

FIGURE 2

Pairwise fixation index (Fst) in P. vivax and P. falciparum populations using SNP and MS. Fst was calculated for P. vivax (A) and P. falciparum (B) using MS (red) or SNP barcode (blue). Pairwise Fst values (points) and 95% confidence intervals (lines) were estimated with 1000 bootstraps using the diveRsity package in R.

FIGURE 3

Clustering analysis by STRUCTURE in P. vivax and P. falciparum populations using SNP and MS. The graph depicts the clustering models when the isolates were assigned into different number of clusters (K). Each isolate is represented by a single vertical line broken into K colored segments, with lengths proportional to each of the K inferred clusters. (A) Clustering of P. vivax populations sorted by location at K = 3 and 4 when using SNP and MS. (B) Clustering of P. falciparum populations sorted by location at K = 4 and 5 when using SNP and MS. (C) Clustering of P. falciparum populations in NJ by previous classification according to (Cabrera-Sosa et al., 2024) at K = 4 and 5 when using SNP and MS.

FIGURE 4

Principal component analysis (PCA) of P. vivax and P. falciparum populations using SNP and MS. PCA was performed for P. vivax (A, B) and P. falciparum (C, D) using MS (A, C) or SNP (B, D). Each community is represented by a unique shape and color. Consistent shapes are used to identify the same community across both P. vivax and P. falciparum populations.