One needs a theory of mechanism to develop an intervention i.e. "new drug". The simpler the better. But given the tremendous complexities involved in biologica

One needs a theory of mechanism to develop an intervention i.e. "new drug". The simpler the better. But given the tremendous complexities involved in biological systems one has to be very careful with the data. Quite frankly most in the field are not sufficiently trained or aren't rigorous enough in dealing with uncertainties and I wouldn't blame them it's mind-boggling and paralysing. This of course can be exploited given the incentives stated above combined with the public's unawareness of e.g. statistical significance vs clinical significance etc., with the myriad mathematically correct ways of presenting data, and a handwaving attitude about reproducibility, this really opens up the floodgates for someone in a lab to become "creative", an entrepreneur. Which of course should be highly discouraged in the context of scientific research and in order to relieve some pressure the community once in a while condemn the most daring of examples without imhv looking too deep into the entrenched mechanisms enabling this. A PhD student (biomedicine) once jokingly told me about the old HeLa contamination "problem"[0] in labs experimenting with cells. I must have reacted shocked as he laughingly added: Well, that was quite a long time ago, the problems now, only compounded. So, in order to get the full picture it is "healthy" to zoom out and look at the other side of the spectrum i.e. longitudinal studies (with its own sets of limitations). One impressive one relating to Alzheimer's is the famous nun study[1]. [0]https://en.m.wikipedia.org/wiki/HeLa#Contamination https://en.m.wikipedia.org/wiki/HeLa#Contamination [1]https://en.m.wikipedia.org/wiki/Nun_Study https://en.m.wikipedia.org/wiki/Nun_Study