Apparent high recombination rates in clonal parasitic organisms due to inappropriate sampling design

Sampling design is of primary importance for empirical studies, in particular, population genetics. For parasitic organisms, a rather frequent way of sampling individuals from local populations is to collect and genotype only one randomly chosen parasite (or isolate) per host individual (or subpopulation), although each host (subpopulation) harbors a set of parasites belonging to the same species (that is, an infrapopulation). Here, we investigate, using simulations, the consequences of such sampling design regarding the estimates of linkage disequilibrium and departure from the Hardy-Weinberg expectations (H-WE) in clonal parasites with an acyclic life cycle. We show that collecting and genotyping only one individual pathogen per host individual (or per subpopulation) and pooling them to form one 'artificial' subpopulation may generate strongly misleading patterns of genetic variations that may lead to false conclusions regarding their reproduction mode. In particular, we show that when subpopulations (or infrapopulations) are genetically differentiated, (i) the level of linkage disequilibrium is significantly reduced and (ii) the departure from the H-WE is strongly modified, sometimes giving a forged picture of a strongly recombining organism despite high levels of clonal reproduction.