A system of methods for analyzing the genetics of congenital heart disease is to be devised. It will be directed to unifying three approaches: genetic statistics; formal modeling adaptable to embryological processes; and empirical analysis of data from experimental embryology and clinical cardiology. In the first place, the empirical data are an adjunct rather than a main area of enquiry: that is, they will be seen as a source of ideas and as a sounding block for new methods. While the immediate concern is not with analyzing the data, our long term goal will be to apply the methods extensively to several sets of data available to us. Thus, at the same time as the theoretical studies are in progress there will be a preliminary inspection and organization of the data on cogenital heart defects. Most of this empirical material is being dealt with at the University of Maryland. The substance of the theoretical development proposed here is (1) To examine models already formulated (e.g. two-locus models with epistasis; age-dependence; threshold processes) from a statistical standpoint i.e. examine the statistical power of tests between competing models, and the small-sample properties of maximum likelihood estimators (MLE) of genetic parameters (bias, consistency, normality, efficiency etc) in detail, mainly by Monte Carlo simulation. (2) To develop new models which will draw on classical genetic ideas of """"""""developmental homeostasis,"""""""" economical and adaptable modeling, and the conclusions of embryology, with the prospect of identifying basic features that provide a subject for study that is more rational and, being metrical, more coherent than the classical anatomical classifications used hitherto. Maximum likelihood estimators for the parameters will then be developed and explored, as before, by simulation. (3) To explore how readily and reliably one could detect plausible classical patterns of inheritance of these parameters.
Showing the most recent 10 out of 41 publications
