Prevalence of infection
In total, 1002 samples were collected from 28 villages, with 9-100 individuals sampled per village (Figure 1, Additional File 1). 900 individuals were surveyed from 9 Pahari groups (99-101 individuals per group) as well as 100 individuals who identified as Bengali. Two individuals from a tenth Pahari group, the Lushai, were sampled in a majority Bawm village. Median age of participants was 31 years (range 5-80). Overall, 59.9% (600/1002) of individuals were female. Two individuals reported travel in the prior 4 weeks, one individual reported fever within the previous 48 hours, and two individuals reported having had malaria in the last 90 days. These numbers were too low to assess them as risk factors for infection.
In total 42/1002 (4.2%) of individuals tested positive by qPCR for P. falciparum and/or P. vivax (21 P. falciparum mono-infection, 16 P. vivax mono-infection, 5 mixed) (Table 1). No P. ovale or P. malariae infections were detected. Only 7/1002 (0.70%) individuals tested positive by RDT or microscopy, one of which was not confirmed by qPCR. RDT and microscopy detected 6/26 P. falciparum infections and 1/5 mixed infections (Table 1). All individuals who tested positive by RDT or microscopy were aged 7 to 20 years.
Plasmodium spp. infections detected by various diagnostic methods
Table 1
Plasmodium spp. infections detected by each diagnostic method. Many more infections were detected by qPCR than by RDT or microscopy. One P. falciparum infection detected by RDT was not confirmed by qPCR.
Diagnostic
|
P. falciparum
|
P. vivax
|
Mixed
|
Total
|
qPCR
|
21
|
16
|
5
|
42
|
RDT
|
6
|
0
|
1
|
7
|
Microscopy
|
3
|
0
|
1
|
4
|
Prevalence by qPCR was highest in individuals aged between 11 and 20 years (7.69%, 16/208), and lowest in individuals older than 60 years (1.25%, 1/80). The only significant difference was between the 11-20 years and 21-30 years (1.62%, 3/185) age groups (ANOVA: p=0.045; Tukey Post Hoc: p=0.043) (Table 2). The prevalence in men was 5.97%, (24/402) compared to 3.00% (18/600) in females (p=0.021). Parasitemia did not differ among age groups for either P. falciparum (p=0.545) or P. vivax (p=0.749).
Age distribution of Plasmodium spp. infections
Table 2
Age distribution of participants with Plasmodium spp. infections. The highest prevalence was in the 11-20 age group (7.69%) and the lowest prevalence was in the >60 age group.
Age (yrs)
|
n
|
Infections
|
% Prevalence
|
5-10
|
98
|
5
|
5.10%
|
11-20
|
208
|
16
|
7.69%
|
21-30
|
185
|
3
|
1.62%
|
31-40
|
207
|
8
|
3.86%
|
41-50
|
143
|
4
|
2.80%
|
51-60
|
81
|
5
|
6.17%
|
>60
|
80
|
1
|
1.25%
|
Total
|
1002
|
42
|
|
There were significant differences in the prevalence of parasitemia between ethnic groups (P<0.0001), with 0% (0/100) of Bengali participants infected and 64% (27/42) infections concentrated among two of the ten Pahari groups (Table 3). In total, 18% (18/100) of Khumi and 9% (9/100) of Mro participants were infected. Among the other 8 ethnic groups, less than 4 individuals had parasitemia detected by qPCR in each group. The differences in ethnicity were not reflected in RDT results. Among 10 Pahari groups, one or two positive RDTs were recorded in five groups.
Distribution of Plasmodium spp. infections among ethnicities
Table 3
Distribution of P. spp. infections among ethnicities as detected by qPCR and RDT. A much higher infection prevalence was detected by qPCR than by RDT. P. vivax and P. falciparum were detected in similar numbers (Pv: n=16, Pf: n=21, Mixed: n=5). Infections were highly clustered, with 64% occurring in the Khumi and Mro groups (27/42). No infections were detected among Bengalis (0/100).
|
qPCR
|
RDT
|
Ethnicity
|
P. vivax
|
P. falciparum
|
Mixed
|
% Positive
|
% Positive
|
Khumi
|
9
|
7
|
2
|
18%
|
1%
|
Mro
|
4
|
4
|
1
|
9%
|
1%
|
Tripura
|
1
|
2
|
1
|
4%
|
0%
|
Tanchangya
|
0
|
3
|
1
|
4%
|
2%
|
Marma
|
0
|
3
|
0
|
3%
|
2%
|
Chak
|
1
|
1
|
0
|
2%
|
0%
|
Chakma
|
1
|
0
|
0
|
1%
|
0%
|
Khyang
|
0
|
1
|
0
|
1%
|
1%
|
Bawm
|
0
|
0
|
0
|
0%
|
0%
|
Lushai
|
0
|
0
|
0
|
0%
|
0%
|
Bengali
|
0
|
0
|
0
|
0%
|
0%
|
Total
|
16
|
21
|
5
|
4.2%
|
0.70%
|
There were significant differences in the prevalence of parasitemia between villages (p<0.0001), with no infections detected in 50% (14/28) of villages. All Mro were sampled in one village (Ampu Para), where prevalence was 9% (9/100). Khumi were sampled in four villages, prevalence was high in two of them: 12% (6/51) in Murungu Bazar, and 36% (12/33) Masim Hostel (Figure 1).
Simple logistic regression identified two ethnicities as significant predictors of infection, Khumi (p<0.0001) and Mro (P=0.001), while village was not a significant predictor (p=0.006).
Parasite genotyping
All 21 individuals with P. vivax parasitemia were successfully genotyped. Among 12 samples successfully typed by amplicon deep sequencing, the median coverage for amplicon marker msp1 was 11342 reads. Genetic diversity was high, with 5-11 alleles per MS marker (mean He=0.81), and 7 alleles for msp1. MOI ranged from 1-3 (median=2.0) and 85% (18/21) were polyclonal.
STRUCTURE plots were generated to infer population structure based on alleles of the major clones in each sample. Plots were generated using k=2-5 as the majority of infection samples came from 2 Pahari groups, with less representation from a few other groups. There was no obvious population structure for P. vivax (Figure 2).
The mean P. vivax IBS within ethnicities was 0.14 for the Khumi (n=11), 0.31 for the Mro (n=5), and 0.29 for Tripura (n=2) (Table 4). No two samples had all 11 alleles in common. Relatedness networks were generated using a minimum threshold of IBS ≥0.5 for each link to generate the theoretical relatedness of meiotic siblings. Only one pair and one triplet of showed IBS>0.5. These small networks included four of the five samples collected from Mro, and one Tripura sample. The Tripura sample was taken from the Hati Bhanga village adjacent to the Mro Ampu village from which all five Mro samples were collected.
Identity-by-state among Plasmodium spp. samples
Table 4
Parasite relatedness within Pahari groups as measured by IBS. For both P. vivax and P. falciparum, mean IBS was greatest among parasites detected among the Mro.
|
Ethnicity
|
n
|
Mean IBS
|
P. vivax
|
Khumi
|
11
|
0.14
|
|
Mro
|
5
|
0.31
|
|
Tripura
|
2
|
0.29
|
|
All P. vivax
|
21
|
0.15
|
|
Ethnicity
|
n
|
Mean IBS
|
P. falciparum
|
Khumi
|
8
|
0.12
|
|
Mro
|
4
|
0.50
|
|
Tripura
|
3
|
0.00
|
|
Tanchangya
|
2
|
0.05
|
|
All P. falciparum
|
20
|
0.11
|
By PCA, P. vivax showed some separation of samples by ethnicity with the Khumi and Mro forming loose clusters, although the Khumi samples appeared evenly throughout most of the plot.
In total 76.9% (20/26) of individuals with P. falciparum parasitemia were successfully genotyped by msp2 size-polymorphic marker and 5 amplicons. The median coverage for amplicon markers was 14,729 reads (range 364-68910, 2.5 and 97.5 percentiles) per amplicon marker and sample. Genetic diversity was high with 14 msp2 alleles (3 FC27 and 11 3D7), and 8-15 alleles per amplicon marker (mean He=0.93 across all amplicon markers). The MOI ranged from 1 to 5 (median=1) and 20% (4/20) of infections were polyclonal. None of the isolates had a pfhrp2 deletion.
As for P. vivax, STRUCTURE analysis, revealed no population structure for P. falciparum (Figure 2), and none of P. falciparum isolates (with haplotypes called at more than 1 marker) had identical haplotypes. The mean IBS within ethnicities was 0.12 for the Khumi (n=8), 0.50 for Mro (n=4), 0.0 for Tripura (n=3), and 0.05 Tanchangya (n=2) (Table 4). Only the P. falciparum from the Mro had a mean IBS of at least 0.5, indicative of close relatedness.
Plasmodium falciparum infections showed a low level of clustering, with 40% (8/20) of isolates linked in a small network including samples from the Khumi (n=3), Mro (n=4), and Tripura (n=1) (Figure 3). Parasite isolates from Khumi individuals in this small network included 2/2 from the northern Murungu Bazar village and 1/6 from the southern village near Masim Hostel 75 km away. The two samples from Murungu Bazar were not linked to each other, but were connected in the network by the sample from Masim Hostel.
By PCA, the P. falciparum haplotypes showed very little separation of samples with samples appearing closely around the origin (Figure 4). All 5 Mro samples clustered fairly close together, although they overlapped with samples from the Khumi, Tripura, and Tanchangya as well.