When lositan is run multiple times on the same data set, there is a large variation in the number of outliers found. For example, for the same data set, we can get 900, 1180, 705 outliers depending on the run. These were all run after shutting down lositan and restarting it.
Would anybody be able to shed any light on this matter?
1 answer
It would be interesting to see the plot to give a better answer, but in general this is perfectly possible: The algorithm is stochastic and some variation is bound to occur (though I have never seen that much).
Try the following:
Large number of sims, click both boxes on the bottom left side.
Also are you using 95 or 99 CIs?
Also, with lots of loci you might want to look at the FDR instead, of the standard CI...
Log in to answer this question.