Population-specific positive selection on low CR1 expression in malaria-endemic regions

Abstract

Complement Receptor Type 1 (CR1) is a malaria-associated gene that encodes a transmembrane receptor of erythrocytes and is crucial for malaria parasite invasion. The expression of CR1 contributes to the rosetting of erythrocytes in the brain bloodstream, causing cerebral malaria, the most severe form of the disease. Here, we study the history of adaptation against malaria by analyzing selection signals in the CR1 gene. We used whole-genome sequencing datasets of 907 healthy individuals from malaria-endemic and non-endemic populations. We detected robust positive selection in populations from the hyperendemic regions of East India and Papua New Guinea. Importantly, we identified a new adaptive variant, rs12034598, which is associated with a slower rate of erythrocyte sedimentation and is linked with a variant associated with low levels of CR1 expression. The combination of the variants likely drives natural selection. In addition, we identified a variant rs3886100 under positive selection in West Africans, which is also related to a low level of CR1 expression in the brain. Our study shows the fine-resolution history of positive selection in the CR1 gene and suggests a population-specific history of CR1 adaptation to malaria. Notably, our novel approach using population genomic analyses allows the identification of protective variants that reduce the risk of malaria infection without the need for patient samples or malaria individual medical records. Our findings contribute to understanding of human adaptation against cerebral malaria.

Fig 1. Geographic locations of the samples included in this study.

The areas colored in blue and orange on the map represent countries where malaria is non-endemic and endemic, respectively, based on a WHO report (World Malaria Report 2018). The 11 population groups analyzed are represented by colored circles. The number in a circle is the number of samples. We used the Tableau v. 2021.2 to create the map images.

Fig 2. Genome-wide percentile ranking of three selection tests.

The standardized iHS [34] (A, B, C, D, and E), PBS [35] (F, G, H, I, and J), and XP-EHH [33] (K, L, and M) values across the CR1 gene region (50kb upstream and downstream) are plotted for the five population groups (three endemic and two non-endemic). The XP-EHH results were the three endemic population groups versus Mongols. The two non-endemic PBS results (I and J) are plotted from the tests for Indian Austroasiatic populations, Mongols, and Europeans. Dots represent SNPs (iHS and XP-EHH) or windows (PBS) having percentile ranking values equal to or lower than top 0.10 (Y axis) over the Mbp position on chromosome 1 (X axis). The CR1 gene and repeat region on the X axis are indicated as green and mesh bars under the plots, respectively. Across the three methods, here we show SNPs with a percentile ranking equal to or lower than 0.1 for the endemic and non-endemic population groups. In addition, green and purple diamonds indicate the SNPs/windows associated with the CR1 expression level (rs2274567, rs3811381, rs3886100, rs11803956, rs12041437, rs17186848, and rs11803366) and erythrocyte sedimentation rate (rs12034598).

Fig 3. A coalescent tree of the CR1 locus.

For the sub-region of the gene locus (intron 20 to intron 27: ~16.7 kb), a coalescent tree was estimated by RELATE [39]. The tree is one estimate out of the 100 replicates, as described in Method. The pink dots on the tree branches represent mutations (SNPs) assigned to the lineages of the tree. Vertical and short bars below the tree correspond to the tips of the trees (each haplotype) of five population groups used to construct the tree. The yellow diamonds indicate the locations of rs2274567 exon 22 SNP and rs12034598 intron 24 SNP on the tree. We detected two distinct haplogroups in the tree, defined by rs2274567, and designated as the L and H haplogroups. The L haplogroup is more frequent in Indian Austroasiatic and Melanesian populations than in Mongols and Europeans. The LS haplogroup is defined by two SNPs, rs2274567 and rs12034598.

Fig 4. Frequency distribution of the LS haplogroup and the malaria mortality rate.

For the ethnic groups living in South and Island Southeast Asia, the haplotype frequency of the LS haplogroup (A and C) and the malaria mortality rate (B and D) are shown on the geographic map. The pie charts show the frequency of the LS haplogroup in each ethnic group with a sample size equal to or greater than 10. The colors in the pie charts represent the population group which the ethnic group belongs to. The mortality rate per 100,000 people is shown as colors in the map, which was retrieved from the malaria atlas project [45]. The haplotype frequencies for all ethnic groups are provided in the S8 Table.