Forum:transition to bioinformatics

Question

[ejms]

I am a biologist who wants to study plant genetics resources to produce useful information for breeders. I have seen that some approaches generate a lot of data and that I will need to know how to handle it if I want to do research that way, which I do. Hence I want to pursue a phd in bioinformatics with a thesis oriented to plant sciences.

What I am concerned about is how likely it is that I will get admitted given that I am not fluent in any programming language and consider myself mostly wet-lab. I will still apply this year, but the doubts are eating me inside. I feel very concerned about it being a wise decision or not. It might be easier to apply for a plant biology phd and try a thesis related to bioinformatics instead of the opposite after all. But I really want this, I really like it. I did my undergrad thesis on transcriptomics and I liked it, I even learned how to script but just a little bit.

I will have to make a thesis for the phd too and will need base knowledge on both the "bio" and the "info" to accomplish it but I need to get into the grad school program first right? So, just how much base knowledge should I have on the "info" part now that I am going to apply? Is it feasible for me to be admitted as I am?

[Medhat]

A: PhD in bioinformatics or biostatistics as a wet-lab scientist?

A: Should I get into bioinformatics

[ejms]

Thanks for the super fast reply, I willl read it!

[Anima Mundi]

Hi, just a quick comment.

But I really want this, I really like it.

Just roll with it ;)

Nowadays there is a strong need for bioinformaticians, if you happen to have a strong interest for bioinformatics and you are fresh from a degree in biology then you are in luck, in my opinion.

[I0110]

It is not too late for you to learn few programming languages. Learn comment line, python and R will certainly help (and they are not difficult). If you do not get in this year, you can add those skills to your CV and you will most likely to get in next year. Besides the "info" part, bioinformatics also needs a lot of statistics. Remember to take statistics courses (inference, regression, Bayesian...), too. Take those for students from statistics major. Otherwise, it's too easy and not useful. Don't think about making a thesis, think about making papers. Eventually, you can just add the papers to your thesis.

Answer 1

[i.sudbery]

About half of my PhD students come to me with no bioinformatics experience. Its totally normal. I'd say do an online course in ONE of python or R. The point here is not to come in with the skills, but to demonstrate a) a commitment to learning b) that its not going to be a massive struggle.

Something simple, like the Codecademy python courses or the Swirl should be fine. Both are free. Neither take that long. Oh, and I second the idea of brushing up your stats. Much of bioinformatics is basically applied statistics, and a student whose statistics are strong is just as likely (if not more likely) to impress me at interview than one who can code. E.g. I always ask candidates "What do you understand by the term p-value?", you'd be surprised how many candidates can't give a satisfactory answer.

Answer 2

[Joe]

My 2 cents as a 50:50 PhD student:

My project started as entirely wet lab, but I've wanted to learn how to code properly for a while, so I pretty much 'invented' some analyses I could perform/questions to ask. So, I jumped in to writing what I considered to be a fairly complicated bit of python code, rather than taking baby steps (I'm not saying this is the right way for all people), and I just learned by doing.

For me, 90% of learning to code is about having a problem to solve and making your life hard in the process of trying to solve it so that you learn something new :P

My thesis now contains 2 experimental and 2 computational chapters, and I've enough confidence in my abilities at this point that I'm happy to apply to bioinformatics based postdoc jobs. You don't necessarily need to do a bioinformatics PhD, and in fact I'd urge you to try and 'be a hybrid'. We need as many bioinformaticians that still have an understanding/appreciation for how wet lab stuff works, as possible IMHO.