Getting started Repeat experiments

We start with a short field trip, that does not reach the outland of statistics yet. If you only have two microarrays and you want to compare them, you proceed as follows: Normalize the data, rank genes according to fold changes, pick from the top of the list as many genes as you like, and acknowledge in your paper that this is an explorative experiment. We believe that no claims on the statistical significance of your findings should be made.

Coming to repeated experiments, our recommendation is: the more the better. Of course, we will not do the benchwork for you and we will not pay for the microarrays. However, we want to raise two important points. First, repetitions are not only for the sake of p-values, they also improve the quality of your list. Without repetitions your list is much worse than it could be. Second, there is almost no statistical analysis without repetitions of experiments. If you want to go beyond explorative research, you need repetitions.

0 0

Post a comment