The proposed research continues in the broad direction of that carried out under the previous proposal, i.e., the problem of making inferences about genetic processes given various genotypic or phenotypic data, but differs from my previous work in emphasizing much more strongly questions in human genetics and, to a lesser extent, in restriction endonuclease data. The work in human genetics continues my current research interests, and proceeds on two levels. First, problems of the general theory of the analysis of data arising from ascertainment sampling schemes will be continued. Second, this general investigation will be specialized to a consideration of the analysis of diseases thought to be linked to HLA, in particular to insulin dependent diabetes (IDDM). Apart from the analysis of this disease using the general theory referred to above, special problems will be considered. These concern the increased susceptibility to IDDM of DR3/4 heterozygotes, resolution of whether IDDM is caused by a locus linked to HLA or by HLA itself together with an unlinked locus and the derivation of the analysis for families of various sizes together with a calculation of the inefficiency of non-optimal approaches to the variable family size problem. The work in restriction endonucleases continues with a question I have pursued in the past. Standard estimation procedures have been shown to be biased when restriction endonuclease data are used and the problem of the necessary and sufficient modelling to correct the bias will be resolved.
