Within-host parasite diversity can arise from sequential bites from mosquitoes that are infected with genetically distinct strains (superinfection) or more commonly from a single mosquito bite containing multiple strains, some of which may be closely related (co-transmission) [
33]. Traditionally, within-host parasite diversity is measured by characterizing a few polymorphic loci to determine the minimum number of genetically distinct parasite strains present (i.e. multiplicity of infections, MOI) (reviewed by Zhong et al. [
33]), but these methods have limited inter-laboratory reproducibility and comparability among studies. SNP panels and analytical tools have been developed to mitigate the limitations of size-polymorphic markers [
15,
34‐
37]. However, SNPs begin to lose resolution in highly complex infections where most loci with reasonable diversity will result in heterozygous calls. Advances in NGS technologies have made it possible to develop highly sensitive targeted deep sequencing approaches [
22,
38‐
43]. These data can provide information not only the number of strains, but also the genetic diversity [e.g. the within-host diversity index (
Fws)] as well as relatedness and genetic structure of parasites in an infection [
22,
43‐
46]. Multi-locus deep sequencing data can provide information on the number and genetic relatedness of strains in an infection, which can be utilized to evaluate spatial differences and temporal changes in transmission. A recent exciting development is the ability to sequence single infected red blood cells, providing high-resolution data, including the ability to understand drivers of genetic diversity within naturally occurring infections and to unambiguously phase genetic data from mixed infections [
33,
47]. However, these methods are not yet easily scalable to the levels required for most epidemiologic work, requiring additional development.
In malaria-endemic regions, multiple infections are common, and within-host diversity indices broadly correlate with endemicity in a non-linear fashion [
14,
28‐
32,
44,
48‐
50]. Reduced within-host diversity has been associated with increased ITN use [
51] and temporal changes in transmission [
50,
52,
53], indicating that theses metrics may be reasonable indicators of changes in transmission intensity. However, these relationships may be more difficult to interpret in areas where infections are not at a steady-state, and in particular may not reflect local transmission intensity, even amongst locally acquired infections, in areas with high rates of importation [
54]. Although NGS technologies are advancing quantification of within-host diversity, they have also introduced new challenges. For example, the depth of sequencing can dramatically affect within-host diversity indices and comparability of findings between studies. Therefore, study design and the sequencing workflow including sample preparation, choice of sequencing platform, sequencing depth and sequence processing pipelines would benefit from standardization across molecular epidemiology studies. Similarly, the sensitivity of computational methods to identify genetic variants and associated errors are crucial for the successful establishment of NGS-based inference of transmission and the possibility to integrate these tools for routine surveillance.