Cross-Species association statistics for genome-wide studies of host and parasite polymorphism data

Uncovering the genes governing host-parasite coevolutionary interactions is of importance for disease management. The increasing availability of host and parasite full genome-data allows for cross-species genome-wide association studies based on the genomic data of in-fected hosts and their associated parasite strains sampled from natural populations ( i.e. natural co-GWAs). Such studies focus on searching for cross-species allelic associations between pairs of host and parasite SNPs. We aim to quantify the power of natural co-GWAs to pinpoint loci under coevolution with respect to the intrinsic complexities of coevolutionary systems, such as the genetic specificity of the interaction and the temporal allele frequency changes resulting from the interaction. Therefore, we develop the cross-species association (CSA) and the cross-species prevalence (CSP) indices, the latter additionally incorporating genomic data from uninfected hosts. To provide an assessment of the statistical power of these indices, we analytically derive their genome-wide False Discovery Rates (FDR) based on the neutral site-frequency spectrum of the host and the parasite population. Using two coevolutionary models, we investigate under which parameter regimes these indices pin-point the coevolving loci. Under trench warfare dynamics, CSA and CSP are very accurate in pinpointing the loci under coevolution, while under arms race dynamics the power is limited especially for gene-for-gene interactions. Furthermore, we reveal that the combination of both indices across time samples is an indicator for the specificity of the interaction. Our results provide novel insights into the power and biological interpretation of natural cross-species association studies.