Data Analysis of Four X-Chromosome Inactivation Experiments

Abstract

Some experiments on mosaics result in data that suggest one or more developmental models usually consisting of alternative biological processes having several unknown parameters. If each experimental subject responded in an identical fashion, a mathematical formulation might be appropriate. However, since random fluctuations are often inherent in the developmental system, a statistical model incorporating randomness is necessary. This paper presents the biological assumptions and a developmental model which appear to be used in data analysis of four experiments (Nesbitt 1971; Fialkow 1973; Nance 1964; and Migeon 1971). The simplest statistical estimation scheme is presented here; complete details are given elsewhere (Kane 1975). The emphasis is on the results of the parameter estimates and on hypothesis testing in the four experiments.