Study site
Samples used in this study were obtained from two major urban cities, Ibadan and Enugu in Nigeria (Fig. 1). Ibadan is in the southwestern part of Nigeria (7.3775° N, 3.9470° E), where malaria is hyperendemic and transmission takes place all year-round [13]. The intense transmission of malaria in Ibadan makes it clinically difficult to distinguish between recrudescence and reinfection after treatment. Enugu on the other hand, is in the southeastern part of Nigeria (6.4584° N, 7.5464° E). There is a high malaria transmission all year-round, with an average malaria incident rate of 15% during dry season and 35% during wet season [14].
Sample collection
This study was part of a larger drug efficacy study investigating the effect of artemether-lumefantrine (AL) and artesunate-amodiaquine (AA) in the treatment of uncomplicated malaria infection. Two drops of finger prick blood was blotted on 3 mm Whatmann filter paper (Whatmann International Limited, Maidstone, UK) before treatment and during follow up on days 3, 7, 14, 21 and 28. However, DNA was extracted from dried blood spot on filter papers obtained before treatment only and was used for this parasite diversity study.
DNA extraction
DNA was extracted from dried blood impregnated filter paper using Qiagen DNA extraction kit and according to the manufacturer’s protocol. Briefly, one-quarter of the dried blood spot was used for extraction and DNA content was eluted in a final volume of 60 μl with buffer AE (Qiagen DNA extraction kit).
Allelic typing of msp2 gene
The polymorphic block 3 of the msp2 gene was amplified by nested PCR following protocols described by Happi and others [4]. PCR reactions were carried out in a final volume of 25 µl using family specific primers and Illustra™ PuReTaq Ready-To-Go PCR Beads (GE Healthcare UK Limited, Little Chalfont Buckinghamshire, Lot 9618045). Polymorphisms on msp2 alone were used for characterization of parasite population structure in patients’ samples because previous studies from Nigeria have shown that it is the most reliable marker for studying P. falciparum population diversity in the country [4, 15,16,17,18]. Two microlitres (2 µl) of the secondary amplification products were resolved by electrophoresis on a 2% Agarose gel and sized against 100 bp molecular weight marker (New England Bio labs, Beverly, MA).
24-SNPs molecular barcode
We initially quantified parasite genomic DNA by qPCR against a standard curve of known concentrations. All quantifications were done in triplicates and P. falciparum HB3 clone genomic DNA (MR4 stock number MRA-155) was used as control. A pre-amplification step was then performed on samples with parasite DNA concentration lower than 0.1 ng/µl [19]. All samples (both pre-amplified and not pre-amplified) were further subjected to genotyping analysis using the 24 SNP-based parasite molecular barcode assay. Barcoding assay was carried out as described by Daniels and others [8].
Genotype determination
The HRM assay was run on the Roche LightCycler 480 Real-Time PCR System using the manufacturer’s settings and the cycling conditions as stated by Daniels and others [8]. The LightCycler 480 Software (release 1.5.1) was also used to analyse the results and make genotyping calls. The reference and alternate alleles of the controls were used to define the SNP of the sample for each of the 24 assays. Samples were called using the melt curve genotyping workflow, in which the software automatically compare melting peaks of the samples to those of the controls (Fig. 2). However, the data was also manually changed in case of any missed, incorrect, or ambiguous genotyping calls. A sample was designated to have mixed (N) infection if peaks that matched two of the control genotypes at a specific locus were observed. A sample was designated to be negative (X) if it showed no discernable peaks or whose peaks did not match those of any of the controls. These genotyping calls were compiled to give the complete 24 SNP barcode for each sample, which was then subjected to further analysis.
Statistical analysis
Data cleaning
Prior to analysis of the raw barcode data, proper clean up measures were taken to ensure that only high quality data were considered for further analysis. Two (2) major parameters were used to clean the data, these were: (i) only ATCG, N (mixed infection) or X (missing SNP) should be present in the barcode data (ii) and samples with barcodes containing more than 4 missing SNPs (X) were excluded from the analysis.
Complexity of infection
Polygenomic infections was established in the parasite population by examining the number of heterozygous SNPs (N) in each sample assayed. It is popular knowledge that the human blood-stage malaria parasites is haploid, therefore, a monogenomic infection should have only one allele at each SNP locus while a polygenomic infection is expected to carry multiple alleles. Therefore, in categorizing the parasite population as polygenomic, a minimum threshold value of at least two (2) heterozygous SNPs (N) was utilized as defined by Sisya and others [20]. This is because a single random SNP out of the 24 SNPs genotyped is occasionally wrongly scored as heterozygous even in well-characterized monogenomic infections. Thus, samples were categorized as monogenomic infections if without any or at most, one (1) heterozygous SNPs (N) in the barcode [20].
Minor allelic frequency
Minor allelic frequency (MAF) was computed according to methods earlier described by Baniecki et al. [12]. Briefly, MAF was calculated from allele counts for each SNP in each population. For each polymorphic genotype calls for both the reference and alternate alleles were calculated by designating each with a half contribution compared to monomorphic genotypes. The average MAF (AMAF) defined as the unweighted mean of the MAF values for both populations for each SNP was further determined.
Population diversity (Barcode, π)
The population diversity (π) was calculated as described by Baniecki et al. [12]. Barcode π is a measure of population diversity with values ranging from 0–1. Values closer to 1 suggest high population diversity [12]. Briefly, π was manually calculated as the mean of the pair-wise differences at assayed SNPs between all members of a population divided by the total number of assayed SNPs. In addition, for each monomorphic/polymorphic genotype, half the original value of the monomorphic/polymorphic mismatch was utilized [12].
Population divergence (fixation index: FST)
The population divergence was measured by calculating the fixation index (FST) for all pairs of populations. The online software, COIL (Broad Institute, USA) was used to determine the FST for both populations.
Principal component analysis (PCA)
Principal component analysis (PCA) was performed with the online program, ClustVis (https://biit.cs.ut.ee/clustvis/) on each of the parasite populations (Ibadan or Enugu) separately as well with both populations together.