Its 4 AM in the morning, but don’t worry, it isn’t a post which I am composing half asleep. I had a virtual espresso shot and somehow just for today- I have become an insomniac. Fun!
I had been working on a funded research project (for last 6 months) focusing on sensitivity analysis and reduced order modeling of mathematical models of infectious diseases like tuberculosis and typhoid. My mentor and I started working on an hypothesis. We strongly believed in the hypothesis as it was highly intuitive and seemed quite practical. Initial experimentation were in-line with the hypothesis we had started with. Gradually the hypothesis got converted into a theory which we wanted to publish. In fact, we had started writing the draft for the Journal of Theoretical Biology or Journal of Mathematical Biology. What went wrong then?
During drafting of the journal paper, one day an idea suddenly struck me (while playing badminton). I had done a semester long course on Machine Learning and its applications in real world problem. Techniques used in machine learning are suitable for analysis of high-dimensional data, something which I was dealing with in my research. I wanted to visualize and cluster my high-dimensional data and provide a strong argument in support of the theory which we were going to present. Ambitious!
We found a technique: t-SNE (t-Stochastic Neighbor Embedding), which worked remarkably well with even very very high (>20000) dimensional data. The crux idea of the technique was fairly simple and we implemented it without any hiccups. To our utter shock, the results from t-SNE shattered the hypothesis to bits-and-pieces. It wasn’t something which my supervisor had asked me to do. It was my own idea which resulted in scrapping of my previous 3 month’s work. Shattering and painful! We simply couldn’t neglect the findings (though we are confident that we would have published the earlier results as well since no where in systems biology, had any one has ever attempted to implement t-SNE algorithm. No one would have challenged us!). It was the matter of upholding research integrity: that as a researcher we have to put the most correct form of information in front of the world.
None the less, there were beautiful lessons to be learnt from this incident. I learned about the strength of integrity and honesty. Staying true and honest to my work was really difficult and required very deep work ethics. I learned that a certain amount of knowledge is never enough! Never! There is always something which you will not be aware of. You really need to work hard to keep up with recent research trends in all the domains which are even remotely connected to your work. And this is very essential as you never know what might become useful for you. Most importantly, I learnt to tackle the failure.
Discussing with my mentors, I realized that what happened with me is not rare. Rather it is quite common and happens with almost everyone at some point in their research. Instead of withdrawing from the problem in fear of further failures, one should take inspiration from it. Use it as a building material for your next hypothesis which would be that much more robust. Catching your own mistakes is no smaller feat either since it is difficult for us to scrutinize our own creation. Yes, there will be a bitter taste left in you after such incidents but that taste will keep you on your toes, ever so cautious and attentive with your next steps.
What did I do in my case? For starters, I didn’t hide anything from my supervisor. I struggled to come up with a new idea/hypothesis to test for a long time. Partially because moving on from a hypothesis I strongly believed in was difficult and partially because systems biology isn’t my long term research domain. Some inertia is justifiable! Currently, I have a new idea which I am eager to discuss with my supervisor. It is something which gives me joy that I can push myself beyond my own limits in challenging circumstances. I have a new beginning at least, which is very important. Incidents like these test your character and work ethics in the most challenging circumstances. I am happy to be able to just pass this test. The entire experience was quite humbling.
Sadly, for all the joy (of solving problems) research gave me, this is one of its dark sides which I got to experience. The wiser people say that: “The more you struggle initially, the easier your path will be become later on”. I hope that is applicable to me as well. Apart from these hiccups, research is beautiful, elegant, powerful and at times eccentric! Really really enjoyable! I hope the readers will get something out of my own experience (which is the prime motivation of writing this blog post).
See you next time. 🙂