We found an average ρ=0.97, although it may be noted that the correlation is slightly lower (ρ=0.907) for the set excluding Finland.
Notice that a random set is intrinsically different from a true negative set: while negative and positive sets should be logically distinct, positive and random sets may overlap regardless of the size of the sets (excluding the trivial scenario of empty positive/random sets).
Note that for clusters marked as "outlier", their quality was estimated from the remainder of the data set, excluding the samples with a proportion of outliers >20%.
(For the cross-participant analyses, the 80 voxels were those that were most stable across the 60 words for the participants in the training set, excluding any data from the participant involved in the test of the classifier).
One data set includes all observed data points within the MC intervals, and the other data set excludes data points corresponding to the Aflvénic fluctuations (14:10–16:50 UT).
The mean Pearson correlation was taken as the mean pairwise correlation between all genes in the gene set, excluding the unit correlation of each gene with itself.
The average error for all validation set compounds and the validation set excluding the 43 compounds were found to be similar (2.29 for all compounds vs 2.34 for compounds that are within the descriptor space).
Results from regression models for the data set excluding the Thames registry are presented in Table 1 and are described here.
To examine the evolutionary persistence of functional compensation for different phenotypes, we compared the ages of duplicate genes among the four phenotypes in either the entire data set or the data set excluding tandem duplicates (fig. 3).
At this point the threshold imposed on the SNP set excluded some Sativa accessions from further clustering (i.e. Pinot Noir, Gewürztraminer, Rhein Riesling, Aromriesling, Sauvignon, Sauvignonasse, Perle and Sacy among the true-to-type individuals).
