In Africa, control programs that target primarily Plasmodium falciparum are inadequate for eliminating malaria. To learn more about prevalence and genetic variability of P. malariae in Africa, we examined blood samples from 663 asymptomatic and 245 symptomatic persons from western Kenya during June-August of 2014 and 2015. P. malariae accounted for 5.3% (35/663) of asymptomatic infections and 3.3% (8/245) of clinical cases. Among asymptomatic persons, 71% (32/45) of P. malariae infections detected by PCR were undetected by microscopy. The low sensitivity of microscopy probably results from the significantly lower parasitemia of P. malariae. Analyses of P. malariae circumsporozoite protein gene sequences revealed high genetic diversity among P. malariae in Africa, but no clear differentiation among geographic populations was observed. Our findings suggest that P. malariae should be included in the malaria elimination strategy in Africa and highlight the need for sensitive and field-applicable methods to identify P. malariae in malaria-endemic areas.

Single nucleotide polymorphisms (SNPs) have been shown to be useful in revealing population structure with continental-and regional-scale samples. In epidemiological study, a careful selection of SNPs to track disease spread in local communities would provide an important addition to traditional disease surveillance. This study used SNPs and microsatellites to examine population structure of Plasmodium falciparum at fine- scale in malaria-endemic areas of Western Kenya. A set of high performance (HP) SNPs were selected from a large SNP panel based on BELS ranking, FST values and minor allele frequency criteria. The discriminative power and assignment accuracy of different SNP panels including nonsynonymous SNPs, silent SNPs, previously published barcode SNPs, and the HP SNPs were evaluated together with microsatellites. Among all SNP panels, HP SNPs showed the highest level of differentiation and self-assignment accuracy on average among sites. Clear distinction was observed between the northern and southern P. falciparum samples, whereas samples from the south were least diverged from one another. These results were comparable to those by microsatellites. Nonsynonymous, silent, and barcode SNPs all showed similar levels of genetic variability to one another and weaker structure than the HP SNPs. We described here the procedure of selecting a set of HP SNPs from a large panel of SNPs that resolve population structure of P. falciparum between the northern and southern regions of Western Kenya. Future work is needed to determine if this procedure can result in SNPs panels capable of tracing Plasmodium spread at finer geographical scales.