The MaxMean summary method first calculates the deviation from the Row Reference Summary Statistic (mean or median) for each sample at each gene. For each individual, all positive values of deviation, and then all negative deviations, are averaged separately across all genes in each gene set defined by the Category Variable. The mean deviation for each gene set that has the largest magnitude is used as the gene set score for each individual. An individual assigned a negative score for a gene set indicates that genes within that set tend to be down-regulated for the individual in reference to the control set level, whereas a positive score means that the individual exhibited more up-regulation in comparison to the reference level.
The MeanAbsolute summary method takes the absolute value of the deviation (from the reference value) for each sample at each gene (row) in the data set. The individual score for each gene set is the average of these absolute deviations across all genes in the gene set. The MeanAbsolute scores are always positive and are computed based on all genes in the gene set (as opposed to MaxMean, which is based only on either genes within the set with negative deviations OR genes in the set with positive deviations), but information as to whether genes in the set tend to be up-regulated or down-regulated in comparison to the reference level is lost.
