Estimação de proporções alélicas e genotípicas individuais de dados CNVs (Copy Number Variations)

Abstract

Copy Number Variation (CNV) are DNA segments that exhibit variations in the number of copies of a sequence (gene), in relation to the usual number of two copies per individual, ranging from one kilobase to three megabases in size. Numerous studies have identified an association of the variations in the number of copies of some genes with various complex genetic diseases such as lupus, rheumatoid arthritis and type 1 diabetes, HIV infection and malaria. Most methods available for identifying CNVs are able to describe only the total number of copies of a gene or DNA segment per individual, leaving behind the number of copies per chromosome. Gaunt et al. (2010) developed a program called CoNVEM, which is based on the EM algorithm (Expectation-Maximization) to determine the allele frequencies in haploids of CNV data, assuming the data are in Hardy-Weinberg equilibrium (HWE). However, when the data are in Hardy-Weinberg Disequilibrium there aren´t any statistical tools to estimate the proportions of allelic and genotypic data CNVs. We propose a method to estimate these proportions when there is inbreeding in the population (imbalance), based on the EM algorithm and the approach of the profiled likelihood function. The calculations have been implemented in a program that we call by CNVice (Inbreeding Coefficients Estimation for CNV data), using R programming language, developed in partnership with the Laboratory of Human Genome Diversity, http://ldgh.com.br/, of Federal University of Minas Gerais.