Application of a new multi-locus variable number tandem repeat analysis (MLVA) scheme for the seasonal investigation of Cryptosporidium parvum cases in Wales and the northwest of England, spring 2022

Abstract

The protozoan Cryptosporidium parvum is an important cause of gastroenteritis in humans and livestock, and cryptosporidiosis outbreaks are common. However, a multi-locus genotyping scheme is not widely adopted. We describe the further development and application of a seven-locus multi-locus variable number of tandem repeats analysis (MLVA) scheme. From 28th March to 31st July 2022, confirmed C. parvum stools (n = 213) from cryptosporidiosis patients (cases) in Wales (n = 95) and the north west of England (n = 118) were tested by MLVA. Typability (defined as alleles identified at all seven loci in a sample) was 81.2% and discriminatory power estimated by Hunter Gaston Discriminatory Index was 0.99. A MLVA profile was constructed from the alleles, expressed in chromosomal order. Profiles were defined as simple (single allele at each locus) or mixed (more than one allele at any locus). A total of 161 MLVA profiles were identified; 13 were mixed, an additional 38 simple profiles contained null records, and 110 were complete simple profiles. A minimum spanning tree was constructed of simple MLVA profiles and those identical at all seven loci defined genetic clusters of cases (here, null records were considered as an allele); 77 cases formed 25 clusters, ranging from two to nine (mode = two) cases. The largest cluster, following epidemiological investigation, signalled a newly-identified outbreak. Two other cases with mixed profiles that contained the outbreak alleles were included in the outbreak investigation. In another epidemiologically-identified outbreak of six initial cases, MLVA detected two additional cases. In a third, small outbreak of three cases, identical MLVA profiles strengthened the microbiological evidence. Review of the performance characteristics of the individual loci and of the seven-locus scheme suggested that two loci might be candidates for review, but a larger dataset over a wider geographical area and longer timeframe will help inform decision-making about the scheme by user laboratories and stakeholders (such as public health agencies). This MLVA scheme is straightforward in use, fast and cheap compared to sequence-based methods, identifies mixed infections, provides an important tool for C. parvum surveillance, and can enhance outbreak investigations and public health action.

Keywords: Cluster; Cryptosporidium parvum; MLVA; Multi-locus; Outbreak; Subtyping.

Graphical abstract

Fig. 1

Illustration of true, artefactual and bleed-through peaks from the SeqStudio Genetic Analyser in BioNumerics software. A True peaks (circled in green) (with peak heights of ≥ 150 RFU) as well as short and wide artefactual peaks (circled in red), which can be observed at the same number of base pairs across different samples, including an NTC (displayed in the bottom image). B A true peak (circled in green) as well as short and narrow stutter peaks (indicated by red arrows). C Bleed through of strong peaks that occur in the same position across other channels. In the green channel, the true peak (highlighted in green) bleeds through to the blue and yellow channels (indicated by red arrows). In the blue channel, the true peak (circled in green) also bleeds through to the green and yellow channels (highlighted with red arrows). In the yellow channel, the true peak (circled in green) also bleeds through to the green channel (indicated by a red arrow).

Fig. 2

Pie graphs illustrating the frequency with which alleles were detected at individual loci, with each segment representing the proportional fraction each allele contributed to all those at each locus. The number of alleles and range of allele frequency (%) for each locus is shown. The colours are used simply to make the segments clear within each pie.

Fig. 3

AMaCAID analysis of the maximum number of genotypes discriminated by optimal marker combinations for the 7-locus and 6-locus schemes. The graphs show the number of MLVA profiles generated by sequential inclusion of optimal loci within each dataset with cgd1 (top) and without cgd1 (bottom). A single representative of each profile was used; samples with null records and mixed alleles were excluded from the dataset prior to this analysis.

Fig. 4

Size distribution of MLVA clusters with simple or mixed MLVA profiles, among 92/213 Cryptosporidium parvum cases. The numbers above the columns are the number of clusters with n cases. The MLVA profiles and notes on the public health significance of clusters are provided in Supplementary file S3.

Fig. 5

Minimum spanning tree of MLVA of 147 simple profiles, displaying non-cluster cases, outbreak cluster cases, and other clusters of cases. Each MLVA profile is indicated by one node or branch tip displayed as circles and connected by branches. The branch line style indicates the number of loci that differ between MLVA profiles: thick solid line, 1 difference; thin solid line, 2 differences; dashed line, 3 differences; no line, ≥ 4 differences. The branch length is not indicative of genetic distance. The wedges in the circles indicate the number of samples with that MLVA profile.