Regions highlighted from the top 200 SNPs according to SNP importances with RF and T-Trees on CDqc.

Regions highlighted from the top 200 SNPs according to SNP importances with RF (top) and T-Trees (bottom) on CDqc. Each row corresponds to a set of SNPs obtained by merging contiguous SNPs in the rankings that are not separated by more than 20 SNPs. For readability, only groups of more than 2 SNPs appear in the tables. Markers that are isolated but reported as associated in [25] are nevertheless compiled at the bottom of both tables (6). For each region, the columns provide the chromosome number, the number of important SNPs in the region, the most important SNP in the region (and its gene name if provided by PheGenI [40]), the p-value of this SNP and its importance. (1) and (2): the regions reported as strongly (with a trend or a genotypic p-value −5) and moderately (with a trend or a genotypic p-value between 10−5 and 10−4) associated in [25]. (5): also reported by [37]. (4): regions identified by both RF and T-Trees. (3): the two novel regions mainly spotted by T-Trees.