trying to use multivariate normal distribution with high covariation for adjacent loci + improving the GWAS assess by using bonferroni correction and actually counting the number of correctly classified/detected QTL