 Last slide, as I stated, I'll say a few seconds ago, is Affymetrix's own algorithm, something that may have helped. Statistics, let's talk about what you do after you've done the quality assessment of your data. The next thing that you do is you apply some statistical analysis. As I mentioned this morning, basically many different types of statistical analysis that you can provide to microdata and being a focus almost exclusively on two sample tests. Simple for simplicity, and also because this is very applicable to a large fraction of our list. This is remote. Anybody want to add to that or change it? So it changes what you're comparing it to. Totally wise error.