Genome-wide Multiple Loci Mapping in Experimental Crosses by the Iterative Adaptive Penalized Regression

Genome-wide multiple loci mapping can be viewed as a variable selection problem where the major objective is to select genetic markers related with a trait of interest. This is a challenging variable selection problem because the number of genetic markers is large (often much larger than the sample size) and there are often strong linkage or linkage disequilibrium between markers. In this paper, we developed two methods for genome-wide multiple loci mapping: the Bayesian adaptive Lasso and the iterative adaptive Lasso. Compared to the existing methods, the advantages of our methods come from the assignment of adaptive weights to different genetic makers, the iterative updating of these adaptive weights, and the ability to penalize most regression coefficients to be exactly zero. We evaluate these two methods as well as several existing methods in the application of genomewide multiple loci mapping in experimental cross. Both large-scale simulation and real data analysis show that the proposed methods have improved variable selection performance. The iterative adaptive Lasso is also computationally much more efficient than the commonly used marginal regression and step-wise regression methods.