Genetic Diversity of Cryptosporidium hominis in a Bangladeshi Community as Revealed by Whole-Genome Sequencing

Abstract

We studied the genetic diversity of Cryptosporidium hominis infections in slum-dwelling infants from Dhaka over a 2-year period. Cryptosporidium hominis infections were common during the monsoon, and were genetically diverse as measured by gp60 genotyping and whole-genome resequencing. Recombination in the parasite was evidenced by the decay of linkage disequilibrium in the genome over <300 bp. Regions of the genome with high levels of polymorphism were also identified. Yet to be determined is if genomic diversity is responsible in part for the high rate of reinfection, seasonality, and varied clinical presentations of cryptosporidiosis in this population.

Figure 1.

The frequency of Cryptosporidium hominis infections was dependent on both the time of year and the parasite genotype. Infants <6 months are infrequently infected with Cryptosporidium and in our study cohort none were infected. The risk of experiencing a Cryptosporidium infection increases thereafter. The graphs therefore, start at the beginning of the second study monsoon (2015) when Cryptosporidium infections began to be observed in the Mirpur cohort who were >6 months of age. The x-axis in both graphs indicates the month and year. A, Rainfall in the Dhaka area. The y-axis indicates total monthly rainfall (mm). B, The number of surveillance stools collected from children 6–24 months in age was used as a surrogate marker for the number of children participating in the study who were susceptible to cryptosporidiosis. The y-axis indicates the percentage of children experiencing genotyped Cryptosporidium infections occurring during this time period (surveillance and diarrheal). Genotypes are indicated by color with the exception of genotypes which occurred with a frequency of <1% in that month’s samples, which are all colored gray. Only 2 genotypes (IaA18R3 and IbA9G3a) were common in both 2015 and 2016. The months March–June 2017 were omitted from the graph as <50 participants of the correct age range remained in the study.

Figure 2.

Genomic diversity of Bangladesh Cryptosporidium hominis. A, Phylogenetic tree of C. hominis genomes. Neighbor-joining tree was based on the high confidence single-nucleotide polymorphisms (SNPs) occurring with a frequency ≥0.2 in the 32 Mirpur genomes. Branches are labeled with genome ID and associated gp60 genotype. Scale bar indicates genome-wide distance between samples in nucleotide substitutions per site. *Genomes from a “paired” oocyst preparation (isolated from different aliquots of the fecal material collected from a single child at one time point [icddr,b 3 and 4] or at a 10-day interval [icddr,b 19 and 20]), indicating that diversity can occur in the parasite population within a single child. B, Cryptosporidium hominis polymorphic regions. Graph indicates the distribution of allele frequencies, and the number of SNPs on the C. hominis chromosomes. Pi (y-axis) is defined as the average number of nucleotide differences per nucleotide site between 2 samples, and is a measure of allelic heterogeneity. The x-axis (V2) indicates the chromosome position (bp). C, High rates of recombination are apparent in this C. hominis population. Average linkage disequilibrium (LD) (y-axis) between neighboring SNPs as a function of the distance (bp) separating them in the genome (x-axis) for Bangladesh C. hominis. The graph was generated using the data from 32 Bangladesh C. hominis genomes, which had >80% genome coverage of at least 10 times. The pairwise values were calculated using the SNPs with minor allele frequency >0.20. The y-axis indicates the average R² value (the square of the correlation coefficient of 2 SNPs) and x-axis the physical separation 0–2000 bp.