Across a wide range of organisms, traits related to performance or fitness such as reproductive success, birth rate, body size, rate of development, survival, and parasite susceptibility have all been linked to measures of individual genetic variability, or heterozygosity (Acevedo-Whitehouse et al., 2005, Buzan et al., 2020, Charpentier et al., 2008, Coltman et al., 1999, Haanes et al., 2013, Ryder et al., 2010, Slate et al., 2000, Slate and Pemberton, 2002). These so-called heterozygosity-fitness correlations (HFCs) can arise for three main reasons. First, in instances where HFCs are driven by processes such as inbreeding, the measure of genetic variability is expected to reflect a reduction in genome-wide heterozygosity and an increase in the expression of deleterious alleles (general effect hypothesis, (Hansson and Westerberg, 2008, Szulkin et al., 2010)). Alternatively, HFCs may occur if the chosen markers are in genes that have direct effects on fitness (direct effect hypothesis, (David, 1998)), or if they are linked with genes under selection (local effect hypothesis, (David, 1998, Hansson and Westerberg, 2008, Lynch et al., 1995)). Because HFCs can reflect patterns of selection in nature, it is essential to understand how frequently they occur and what mechanisms drive observed patterns in natural populations.

While our understanding of the underlying mechanisms causing HFCs has improved (Acevedo-Whitehouse et al., 2005, Brambilla et al., 2018, Mitchell et al., 2017a, Portanier et al., 2019, Szulkin et al., 2010), the biological significance of these patterns is still under intense debate, in part, because the detectability of HFCs is notoriously inconsistent (Hulse et al., 2023, Martin et al., 2021). For example, a meta-analysis of HFCs in wild and domestic animal populations found that although correlations arise frequently, the associations tend to be weak and often differ from one population or taxon to the next (Chapman et al., 2009). The authors of the study suggested that discrepancies in detecting HFCs might relate to the context under which individuals are sampled, such as differences in the environment. In support of this hypothesis, several studies suggest that environmental context is an important determinant of the strength and detectability of HFCs, however, the type of environment necessary to detect associations is highly inconsistent. For example, in some instances HFCs were only detected under stressful conditions (e.g. limited resources, habitat disturbance) (Brock et al., 2015, Ferrer et al., 2016, Lesbarrères et al., 2005, Marr et al., 2006), while in others, HFCs were observed when conditions were favorable (Annavi et al., 2014, Harrison et al., 2011). In addition to environmental context, non-genetic characteristics of individuals, or life history context, might also influence the detectability of HFCs. For example, some studies suggest that HFCs might only be detected for individuals of a certain age (Judson et al., 2018) or sex (Arct et al., 2017, Rioux-Paquette et al., 2011). Study design can also affect detectability of HFCs, if for example, the sample size is small or few genetic markers are used to quantify heterozygosity (Chapman et al., 2009). These findings emphasize that a focus on both environmental and life history context may be central to understanding which mechanisms influence the occurrence of HFCs in natural populations, and that study design may also be influential in detecting patterns.

Context-dependency and study design might be particularly important for understanding heterozygosity-parasite correlations (HPCs). Parasites often negatively impact the fitness of their hosts by reducing growth and development rates, body mass, condition, survival and reproduction (Hillegass et al., 2010, Hurtrez-Boussès et al., 1997, Sperry et al., 2009, Ujvari and Madsen, 2006), and as such, have been the target of many HFC studies (e.g., Acevedo-Whitehouse et al., 2003, Coltman et al., 1999, Ferrer et al., 2016, Hawley et al., 2005, Rijks et al., 2008, Voegeli et al., 2012). However, variable outcomes are common in HPC studies. For example, across a range of host-parasite systems, some studies report no correlation between heterozygosity and different metrics of parasitism (Kubacka et al., 2020, Vallender et al., 2012), whereas other studies report negative (e.g., Brambilla et al., 2018, Budischak et al., 2023, Ferrer et al., 2014, Hoffman et al., 2014, Mitchell et al., 2017b, Townsend et al., 2018) or positive (e.g., Ferrer et al., 2016, Sutton et al., 2016) correlations. Indeed, even within the same host population, HPCs have been reported for some parasites but not others (Charpentier et al., 2008, Ruiz-López et al., 2012a, Townsend et al., 2018). Given that fitness consequences of infection can vary depending on traits of the parasite (Sweeny et al., 2022) and host (Gkafas et al., 2020, Shaner et al., 2013), as well as dynamically over time and across environmental conditions (Lesbarrères et al., 2005, Sweeny et al., 2022), fluctuations in the strength of HPCs in wild populations may be the norm rather than the exception.

In this study, we tested for context-dependent HPCs using three complementary study designs (cross-sectional, longitudinal, and experimental) that allowed us to simultaneously examine the occurrence, persistence, and causal basis of HPCs. Focusing on a wild population of Grant’s gazelles (Nanger granti) in Central Kenya and a group of endoparasites (helminths, protozoa) that commonly occur in this population (Ezenwa, 2003, Ezenwa et al., 2012), we first tested for the occurrence of HPCs in gazelles using samples collected from 103 individuals sampled at a single time point (cross-sectional analysis). Next, we examined the persistence of these correlations through time using data from 25 individuals that were sampled repeatedly for parasites over a 12-month period (longitudinal analysis). In both cases, we examined whether HPCs depended on host life-history (sex, age) and environmental (seasonality) context. Finally, we evaluated whether there was a causal link between individual heterozygosity and parasitism by testing for an effect of individual heterozygosity on the propensity for individuals to re-acquire parasites after experimental clearance using data from 15 individuals (experimental analysis). In general, we predicted that individual heterozygosity would correlate negatively with parasite infection, with more genetically diverse individuals hosting fewer parasites or showing slower rates of parasite re-accumulation. However, we also expected that the strength of these relationships would vary depending on host and environmental context as well as study design.