Abstract
Pathogen traits, such as the virulence and the transmissibility of an infection, can vary significantly between patients. A major challenge is to measure the extent to which the genetic differences between infecting strains explain the observed variation of the trait. This is quantified by the so-called broad-sense heritability H2 – a term borrowed from quantitative genetics of sexual species. A recent discrepancy between estimates of HIV-virulence-heritability has opened a debate on the accuracy of the estimators. Here, we show that the discrepancy originates from model limitations and important lifecycle differences between sexually reproducing organisms and transmittable pathogens. In particular, current quantitative genetics methods are prone to underestimate H2, because they do not account for rapid within-host mutation combined with natural selection on the trait. We introduce two independent approaches correcting these errors: ANOVA-CPP and our POUMM method. Empirical analyses reveal that at least 20% of the variation in virulence is explained by the virus genome both for European and African data. These results should terminate the ongoing discussion whether the virus affects virulence at all, and should motivate further genome-wide association studies on the virus, as well as studies on the interaction between host- and viral factors for virulence. Beyond HIV, we discuss that ANOVA-CPP is ideal for slowly evolving protozoa, bacteria and DNA-viruses, while POUMM is ideal for rapidly evolving RNA-viruses, thus, enabling heritability estimation for a broad range of pathogens.
Significance statement Pathogen traits, such as the virulence of an infection, can vary tremendously between patients. To what extent the pathogen rather than the host determines these traits remains a mystery for many infectious diseases. This is quantified by the traits’ “heritability” – a term borrowed from quantitative genetics of sexual species. Recently, a discrepancy between studies of HIV-virulence has opened a debate on the appropriate estimators of pathogen trait heritability. We find the origin of this discrepancy in the inability of current quantitative genetics methods to account for rapid pathogen mutation in combination with natural selection on the trait. We introduce two independent approaches correcting these errors and report agreeing heritability estimates of these approaches on synthetic and empirical HIV-data.