Modelling the pathogenesis of cystic fibrosis and other monogenic conditions, and the occurrence of causative variants

Abstract

Deleterious recessive monogenic autosomal conditions are modelled both on an individual level, for diagnostic purposes, as well as in large populations, where the establishment, dispersion and equilibrium behaviour is investigated.

Data fusion techniques are applied to combine diagnostic data on a more rigorous basis, to support the diagnosis of disease in an individual. In this case the focus is specifically on cystic fibrosis, which is one of the most common monogenic recessive disorders in humans. Diagnostic information may be of disparate types and varying verisimilitude, such as symptoms, measurements, history, observations, and even opinions. Nonetheless it is possible to construct a mathematical framework to synthesise this knowledge into a numeric assessment of the probability that the disease may be present. This may be used to guide decisions regarding treatment or additional testing, by supporting improved cost-benefit analyses.

Considering the population genetics of monogenic variations such as cystic fibrosis, analytical and statistical stochastic approaches are used to model and predict the dispersion of mutations through a large population. These approaches are used to quantify the magnitude of a heterozygous selective advantage of a mutation in the presence of a homozygous disadvantage. Random effects such as genetic drift are accounted for, which are likely to extinguish even highly advantageous mutations while the prevalence is still low. Dunbar’s results regarding the cognitive upper limit of the number of stable social relationships that humans can maintain are used to determine a realistic community size - a reduced local subset of the total population - from which an individual can select mates. This reduction has a dramatic effect on the probability of establishing mutations, as well as the eventual equilibrium values that are reached in the case of mutations conferring a heterozygous selective advantage, but a homozygous disadvantage, as in the case of cystic fibrosis and sickle cell disease. The magnitude of this selective advantage can then be estimated based on observed occurrence levels of a specific mutation in a population, without requiring prior information regarding its phenotypic manifestation.

It is also demonstrated that the heterozygous carrier levels of monogenic recessive disorders are routinely overestimated.