Identification of complex Plasmodium falciparum genetic backgrounds circulating in Africa: a multicountry genomic epidemiology analysis

Abstract

Background: The population structure of the malaria parasite Plasmodium falciparum can reveal underlying adaptive evolutionary processes. Selective pressures to maintain complex genetic backgrounds can encourage inbreeding, producing distinct parasite clusters identifiable by population structure analyses.

Methods: We analysed population structure in 3783 P falciparum genomes from 21 countries across Africa, provided by the MalariaGEN Pf7 dataset. We used Principal Coordinate Analysis to cluster parasites, identity by descent (IBD) methods to identify genomic regions shared by cluster members, and linkage analyses to establish their co-inheritance patterns. Structural variants were reconstructed by de novo assembly and verified by long-read sequencing.

Findings: We identified a strongly differentiated cluster of parasites, named AF1, comprising 47 (1·2%) of 3783 samples analysed, distributed over 13 countries across Africa, at locations over 7000 km apart. Members of this cluster share a complex genetic background, consisting of up to 23 loci harbouring many highly differentiated variants, rarely observed outside the cluster. IBD analyses revealed common ancestry at these loci, irrespective of sampling location. Outside the shared loci, however, AF1 members appear to outbreed with sympatric parasites. The AF1 differentiated variants comprise structural variations, including a gene conversion involving the dblmsp and dblmsp2 genes, and numerous single nucleotide polymorphisms. Several of the genes harbouring these mutations are functionally related, often involved in interactions with red blood cells including invasion, egress, and erythrocyte antigen export.

Interpretation: We propose that AF1 parasites have adapted to some unidentified evolutionary niche, probably involving interactions with host erythrocytes. This adaptation involves a complex compendium of interacting variants that are rarely observed in Africa, which remains mostly intact despite recombination events. The term cryptotype was used to describe a common background interspersed with genomic regions of local origin.

Funding: Bill & Melinda Gates Foundation.

Figure 1

PCoA of African samples, revealing population structure A plot of PC2 versus PC1 is shown. Along PC1 (explaining 1·9% of variance), samples separate geographically so that the east Africa, central Africa, and west Africa macroregions can be distinguished as labelled. A cluster of AF1 parasites, originating from multiple countries, separates along PC2 (0·9% of variance). Two horizontal dotted lines indicate the thresholds for defining the AF1 population. Samples with PC2 of more than 0·025 were classified as AF1; those with PC2 of less than 0·01 were classified as non-AF1; the remaining parasites were disregarded in further analysis, because their AF1 membership status is inconclusive. PCoA=principal coordinate analysis. PC1=first principal component. PC2=second principal component.

Figure 2

Geographical distribution of AF1 parasite samples In the map shown, countries from which parasites were sampled are shown with a coloured background and a label showing the country name. Countries where AF1 parasites were found are shown with an orange background. For each of the countries where AF1 parasites were found, the number of AF1 samples and the total number of analysed samples are separated by a solidus, and the percentage of AF1 samples is shown in brackets. The map uses data from Natural Earth.

Figure 3

AF1 characteristic loci (A) The circular plot maps all 14 nuclear chromosomes (starting clockwise from the top, each chromosome is represented by a coloured segment in the outer ring). The inner ring shows a plot of mean F_ST between AF1 and the three African macroregions (west Africa, central Africa, and east Africa) at non-synonymous coding SNPs (appendix p 7). In high-IBD regions, at least 50% of all AF1 sample pairs are in IBD (appendix p 20). Internal lines show the r² measure of linkage disequilibrium between pairs of high-F_ST SNPs (F_ST>0·2), estimated using all African parasites. Three types of line represent three linkage disequilibrium ranges: r² greater than or equal to 0·2, but less than 0·4; r² greater than or equal to 0·4, but less than 0·5; and r² greater than or equal to 0·5. (B) Presence of characteristic haplotypes in AF1 parasites. This panel shows a matrix of genotypes at each of the SNPs with the highest F_ST in the 23 high-IBD regions identified in the AF1 population. Each row represents an AF1 sample; the sample identification number and the country of provenance are shown. (C) Genes at AF1 characteristic loci. This panel shows maps of gene positions for the ten highest-ranked high-IBD regions identified in the AF1 population. The x-axis represents positions on the high-IBD region’s chromosome. Each gene in the region is shown by a rectangle, labelled with the gene’s name and coloured according to its function (when function is known). The highest-F_ST SNP in each region is detailed in the appendix (p 14). IBD=identity by descent. RBC=red blood cell. SNP=single nucleotide polymorphism.

Figure 3

AF1 characteristic loci (A) The circular plot maps all 14 nuclear chromosomes (starting clockwise from the top, each chromosome is represented by a coloured segment in the outer ring). The inner ring shows a plot of mean F_ST between AF1 and the three African macroregions (west Africa, central Africa, and east Africa) at non-synonymous coding SNPs (appendix p 7). In high-IBD regions, at least 50% of all AF1 sample pairs are in IBD (appendix p 20). Internal lines show the r² measure of linkage disequilibrium between pairs of high-F_ST SNPs (F_ST>0·2), estimated using all African parasites. Three types of line represent three linkage disequilibrium ranges: r² greater than or equal to 0·2, but less than 0·4; r² greater than or equal to 0·4, but less than 0·5; and r² greater than or equal to 0·5. (B) Presence of characteristic haplotypes in AF1 parasites. This panel shows a matrix of genotypes at each of the SNPs with the highest F_ST in the 23 high-IBD regions identified in the AF1 population. Each row represents an AF1 sample; the sample identification number and the country of provenance are shown. (C) Genes at AF1 characteristic loci. This panel shows maps of gene positions for the ten highest-ranked high-IBD regions identified in the AF1 population. The x-axis represents positions on the high-IBD region’s chromosome. Each gene in the region is shown by a rectangle, labelled with the gene’s name and coloured according to its function (when function is known). The highest-F_ST SNP in each region is detailed in the appendix (p 14). IBD=identity by descent. RBC=red blood cell. SNP=single nucleotide polymorphism.

Figure 3

AF1 characteristic loci (A) The circular plot maps all 14 nuclear chromosomes (starting clockwise from the top, each chromosome is represented by a coloured segment in the outer ring). The inner ring shows a plot of mean F_ST between AF1 and the three African macroregions (west Africa, central Africa, and east Africa) at non-synonymous coding SNPs (appendix p 7). In high-IBD regions, at least 50% of all AF1 sample pairs are in IBD (appendix p 20). Internal lines show the r² measure of linkage disequilibrium between pairs of high-F_ST SNPs (F_ST>0·2), estimated using all African parasites. Three types of line represent three linkage disequilibrium ranges: r² greater than or equal to 0·2, but less than 0·4; r² greater than or equal to 0·4, but less than 0·5; and r² greater than or equal to 0·5. (B) Presence of characteristic haplotypes in AF1 parasites. This panel shows a matrix of genotypes at each of the SNPs with the highest F_ST in the 23 high-IBD regions identified in the AF1 population. Each row represents an AF1 sample; the sample identification number and the country of provenance are shown. (C) Genes at AF1 characteristic loci. This panel shows maps of gene positions for the ten highest-ranked high-IBD regions identified in the AF1 population. The x-axis represents positions on the high-IBD region’s chromosome. Each gene in the region is shown by a rectangle, labelled with the gene’s name and coloured according to its function (when function is known). The highest-F_ST SNP in each region is detailed in the appendix (p 14). IBD=identity by descent. RBC=red blood cell. SNP=single nucleotide polymorphism.

Figure 4

DBLMSP gene sequence crossover in AF1 parasites (A) Schematic of the gene conversion underpinning the AF1 variant of dblmsp. The diagram shows as colour blocks the sequences of dblmsp and dblmsp2 in four Plasmodium falciparum genomes: Pf3D7 (reference), PfIT (long-read sequenced), PfKH02 (long-read sequenced), and AF1 (de novo assembly of sample PM0293-C). Blocks of the same colour indicate highly similar (near-identical) sequences. Coordinates shown (not to scale) correspond to the Pf3D7 positions in dblmsp2 (above) and dblmsp (below). The AF1 dblmsp sequence is near-identical to that of PfIT dblmsp2 at the 5′ end, and of PfIT dblmsp after position 991. The AF1 dblmsp2 sequence, however, is near-identical to the dblmsp2 sequence of PfKH02. The grey region is a 19-nucleotide sequence identical in dblmsp and dblmsp2, providing a recombination breakpoint. (B) Detail of the AF1 dblmsp and dblmsp2 breakpoint region. This panel shows an alignment of the AF1 dblmsp sequence (middle) against the dblmsp2 (above) and dblmsp (below) sequences of PfIT. The 19-nucleotide region of 100% identity is shown in green; to the left, the AF1 sequence is identical to PfIT dblmsp2, whereas to the right it is identical to PfIT dblmsp. The underlined 62-nucleotide portion of the AF1 sequence was used as search query to confirm the presence of the conversion breakpoint in the AF1 parasites.