Forum:Transitioning from Wet Lab to Computational Biology as a PhD Student/Surgeon

Question

[ian.folkert]

Hey everyone,

I'm a surgery resident that decided to take some time during my training (now ~4 years of time) to go to graduate school and get a PhD in cell and molecular biology. When I started I knew I was interested in oncology/surgical oncology, but was mostly clueless about what kind of research I wanted to do and hadn't been in a lab since before medical school. I ended up joining a tumor immunology lab, and have spent most of the past 4 years doing the typical wet bench molecular biology, flow cytometry, etc. I took a RNAseq class here to learn how to analyze a dataset I generated as part of my thesis research, and seeing the power of the technology and also how limiting it was to have to rely on others to do even basic exploratory analysis, I became really interested in learning more. I now have some scRNAseq data as well that I am trying to learn how to analyze using resources online, and it has been really fun trying to learn more.

I guess my question is, for someone in my position, what would you prioritize learning? If I could go back in time I would have learned computer science and done a computational PhD, but with a year or two left in graduate school I want to make the most of my time. My goal is to eventually use my access to human patients in the clinic as a surgeon to be able to generate high dimensional data and ask questions using primary human samples. My knowledge of statistics is definitely lacking and a weak spot for me. But, I was wondering if there are other resources or areas of focus people would recommend? Thanks in advance for your thoughts!

Answer 1

[jordi.planells]

Hi Ian. This is a really difficult question to answer. I did my bachelor in molecular biology. Then I did my master in Biotechnology. Now I am doing a PhD in molecular biology. The project is much about chromatin and I am analyzing a lot of high-throughput data (RNA-seq, ATAC-seq, ChIP-seq, etc.) which is pretty much what you would like to do. In my experience, I would invest some time in bash fundamentals (there are many good tutorials out there). Almost all the programs used in any pipeline are command-line based. Moreover, you will most likely work with many samples and knowing how to loop through them will save you a lot of time. Sed, grep and awk are basic commands one should be quite familiar with. In the beginning, I would chose a programming language and commit to it. I would recommend R, as there are many really useful packages for analyzing your data. Regarding the statistical background, it will come as working. You will need to read and chose between different algorithms, go through packages/program documentations, etc. You basically will learn what logistic regressions are, multivariate analysis, covariances, on the fly.

Definitely, the most important thing is that you will learn more by doing. You can be jumping from tutorial to tutorial forever. However, to my experience, when you learn the most is when you challenge yourself and need to do something by your own.

Good luck in your journey!

Jordi

[Carlo Yague]

Definitely, the most important thing is that you will learn more by doing.

--> +1 to this. And since you already have your own RNA-seq dataset to analyse, you are good to go !

[ian.folkert]

Thanks so much for your thoughts! I really appreciate it. I had been trying to learn through tutorials before I had my own data and it was tough in some ways. Having something to "work on" has really helped push me to learn more, so your point definitely seems spot on. I guess I just sometimes doubt whether I will ever be able to get good at this, but I'm trying to just focus on putting in the time each day. Thanks again for your advice.

[GenoMax]

I guess I just sometimes doubt whether I will ever be able to get good at this

You don't have to, since I assume you will be a surgeon first. You just need to understand how some of these things work, if you end up being a physician scientist.