I am investigating the distribution of GSVA scores for different treatment effects. I simulated data with a mean expression value of 2 and added different effect sizes (0.01,0.1,0.2,0.5,1,2,5,10,20,50,100) for the genes in the gene set. Typically, GSVA scores lie between -1 and 1.
with my simulated data, there is an initial decrease in the mean baseline scores as the treatment effect increases. After an added treatment effect of 5, the mean scores start to gradually rise and eventually stabilize around -0.5.
The on-treatment scores mean initially increases with rising treatment. After a treatment effect of 5, the mean scores display a mild decrease, eventually stabilizing around 0.5.
The mean absolute change from baseline (mean | on-treatment scores - baseline scores | ) mirrors the pattern of the on-treatment scores mean but plateaus at an average value of 1.
You can check the photo of the plots here (https://ibb.co/ynw4wVM)
In all cases, the distribution offshoots the limits. Is there any explanation?
