Furthermore, gain-of-function research doesn't have to be by genetic engineering. The controversial experiment of a few years back which I vaguely remember was more like natural selection.