Genetic Data Generation Method Based on Multivariate Binary Random Variables

With the development of society and economy, statistics is increasingly used as an instrumental discipline in our society. Among them, biostatistics is an emerging branch of statistics, which has much room for research. As the basis for all biostatistical researches, biological genetic data play an extremely important role in biostatistics. In previous researches, there have been many methods to generate data. The key to generate genetic data is to ensure the correlation between genetic loci (loci is the plural of locus). However, many researchers ignore the important issue of loci correlation when simulating data, resulting in data generated for each locus being independent and unrelated. Therefore, we used multivariate binary random variables to generate gene loci data (MBRV-data). To verify that our method can be used to generate data for biostatistical studies, we use extensive simulations to determine the feasibility of this method. We also validate the validity of the generated data by investigating population stratification methods. The result of these simulations shows that the MBRV-data can be well applied to the simulation experiment of genotype data. Whether it is the type I error rates or the power, it can reflect the validity and rationality of the method proposed in this paper. This method has been implemented in R.