- You can just keep doing an experiment over and over again and you keep getting the same answer. You know that's good, you're reproducible but at some point, you're not learning anything new. Then that experimental design, and my thesis advisor David Botstein, who said that's "taking rigor "and making it into rigor mortis." Right and so it's a really good statement, right? At some point it's diminishing returns. One way, as experimental design, is to certainly design experiments which test ideas by different approaches.

- So you know, you could do an experiment a 1000 times and optimize that experiment. But unless you find a different way to validate that experiment. I don't think you can ever really, really, it can never be as convincing as if you find a different way to address that same question. For example, let's say you had three different types of experiments. If you had three experiments, all giving you the same answer versus 800 trials of a single experiment. I can tell you which one I think is more convincing. And it's going to be like three independent trials.

- So the nice thing with our orthogonal approach is when they align with one another and you test in your experiment what the outcome was, you have two completely different ways of gauging success that's comforting. I don't know any other way to phrase it other than I would like to know that that experiment was designed to get an answer and if I test the answer to that question in two different ways, and get basically the same confirmation. To me, that suggests that it was a successful experiment.

- I think the other thing that's incredibly powerful about orthogonal approach is that, one it allows you to potentially include different controls because of the nature of the experiment. So that you can feel much more comfortable about the conclusions you're drawing and the caveats you're addressing. And two, it also has the opportunity, to potentially reveal new biology. Because you're doing an experiment in a slightly different way there are things that could buttress your initial results, but also provide new information that tells you that you should be actually going in a different direction as well. So when we are discussing that tension between doing enough experiments to draw confident conclusions versus the amount of time that it takes to do that many experiments Ultimately, you have to be comfortable with the conclusion you're drawing.

- Really good exercise for a lot of people, is to go back and read the famous Hershey-Chase experiment. Which supposedly show that DNA was the hereditary material. So the way the experiment is done. Is you made, viruses that had S35 labeled proteins and P32 labeled DNA. You absorb the virus to the cells and if it's the protein is the genetic material, you'll see the S35 has to go inside the cell. If it's the DNA it'll be the phosphorylation... The phosphorylated signal would go inside. So it's reported out that they did the experiment and the S35 stuck on the outside, so the protein stayed outside, and the phosphorylation went in, so the DNA went in. If you look at that experiment 70% of the phosphorylated DNA went in or stayed within and 30% of the protein. So you say, how could you possibly make this conclusion that they show that DNA is hereditary material. How you do know it wasn't the 30% of the protein? So if you read the paper, that is one of six experiments that they do. Each experiment is slightly different in their approach. So you can't make a model that makes protein by all the other approaches. It's a series of arguments, in which the only that fits all the five different puzzles is the DNA and not the protein. It's really a total misconception and misrepresentation in science that a single experiment is usually ever done, that proves something. So you know, looking to do an orthogonal approach that actually gives you the same test but they get the same answer is a much better way of thinking about experimental design.

- So orthogonal approach could be one of many strategies, but the idea is to come up with either an existing, a different kind of experiment or a different kind of measurement. A certainly fair orthogonal strategy would be using a different measurement technique. So essentially getting a different kind of read out of your biological system. So that, whatever limitations there are about the technique you're using or the blind spots of a technique you know, that could be overcome by a different technique that has different strengths and weaknesses.

- What you really want to do is evaluate what are the pros and cons of each different type of technique. So it's not, there's never necessarily going to be a one silver bullet that is going to give you the best answer using a single technique. So it's good to sort of keep in mind, that you plan to have sort of a multi pronged effort. You want to have multiple ways of testing to see if something is true. Because you may want to take the strengths of one technique and the strength of another technique and combine them. And if two different sort of ways of getting an answer can give you the same results, that is something that is really really powerful.

- Certainly any kind of different kind of experiment is also, I would consider an orthogonal approach. You know basically, a different experimental strategy. Either for perturbing the system in some way or another way of probing that process using a different kind of experimental strategy. And another orthogonal approach also may be to use a different system. That could be in this case for example, a different organism. You know if this phenomena you're studying, or idea that you have may be universal you be not be limited to HeLa cells. You may find it in worms, or you may find it in a different experimental organism. So I think, the answer is probably all of those strategies are good and the more that you can bring them into play in trying to answer biological question, probably the more convincing your study is going to be. The more people will see it as robust, and also potentially applicable to kind of many biological systems.