deletions | additions       diff --git a/section_Discussion_begin_enumerate_item__.tex b/section_Discussion_begin_enumerate_item__.tex  index e63920b..c8680d4 100644  --- a/section_Discussion_begin_enumerate_item__.tex  +++ b/section_Discussion_begin_enumerate_item__.tex 
...
\item Evaluate the localization results from the simulation study.  \begin{itemize}  \item In simulation of a haploid population, Burkett et al. found that haplotype clustering methods based on the true trees outperformed all other association methods in detecting the association signal with a candidate genomic region.  \item Interestingly, in our simulations of a diploid population, the naive Mantel tests performed very poorly relative to the other methods. In fact, Blossoc, CAVIARBF, C-alpha, and Fisher's exact test performed better in localizing the signal.  The poor performance of the naive Mantel test could be explained by the misclassification og haplotypes in the naive Mantel test. In the naive Mantel test, we do not know which of the two haplotypes in a case carries the disease risk variant. When only one of the two haplotypes in a case carries a causal variant, this way of classifying haplotypes could lead to the poor performance. %\item Interestingly, performances were poor for the naive Mantel test. This could be explained due to the misclassification of haplotypes in naive Mantel test. In naive Mantel test, the case status were assigned to the individuals. But, we would not know which haplotype carries disease risk variant among the two haplotypes of an individual. When one of the two haplotypes carried a casual variant, this way of classification could lead to the poor performance.   \item Not surprisingly, the informed Mantel test, outperformed all the other methods since case haplotypes that carry risk variants are grouped together.   % \item Not surprisingly, informed-Mantel test, the bench mark outperformed all the other methods since we cluster the haplotype comes from a case that carry risk variants in informed-Mantel test. However, Blossoc, CAVIARBF, C-alpha and Fisher's exact test performed comparably well in localizing the signal.