want to do methods oriented postdoc from comp bio phd

Question

[phdscm]

Hi,

I'm almost done my PhD in a comp bio program. I have been doing research that I think is cool in terms of scientific value, and has been my own independent work in computational biology, but is not really mathematically/machine learning oriented. I really want to do a postdoc in a more rigorous stats machine learning lab (still in comp bio, but under someone from the cs world) and move my research more in that direction. My undergrad was in cs but not machine learning, and my grad coursework was not very useful. I'm kind of consumed with insecurity that my math/cs background is not enough to be taken seriously by those people. I've been teaching myself from a statistical inference text, and want to follow it with studying cs229, graphical models, optimization. But all this is going to take some time, and obviously I've got a lot else on my plate as I'm defending. Is it better to learn all this before applying to places like that, so that I'm not ruled out as a candidate? Or do you think my "domain knowledge" of computational biology could make up for this weakness? Should I try to get a postdoc somewhere less math heavy and bring those skills into my work on my own? Also if there's a better place to post this q, please let me know.

[Giovanni M Dall'Olio]

There are some very good machine learning online courses in Coursera, they are a good way to learn advanced stats methods.

Answer 1

[russhh]

Personally I'd just apply and be honest about not being an expert within stats/ML. If you can explain which specific methods/areas of machine learning you are concentrating on, and why you feel they are the most relevant to biofx and if you can point to a couple of papers that made you realise why you wanted to move from one area of computational biology to another, all the better. You'll end up learning things continuously as you progress in your career, anyway, so you might as well be explicit about where your technical / theoretical skills could be improved.

Let the PI decide who's right for the job. That you're defending your PhD shows that you've the perseverance to pick up difficult concepts and make things work.

Then again, you might want to wait for someone more senior to reply, than taking my advice without question.

All the best

Russ