How does a physics phd approach bioinformatics with a focus on machine learning?

Question

[brc]

Hi everyone,

I'm currently working as a postdoc (physics) and my main expertise are data analysis, numerical simulation, statistical modeling etc. I mainly use python but I'm also familiar with R. I got interested in the application of machine learning in Bioinformatics, more specifically single cell omics and rna-seq data. I would like to change my research field to single cell analysis. Currently I'm applying for post-doc positions in singe cell data analysis. It will be helpful if you kindly share your thoughts on the career transitions: Do you think if it is ok, doable and practical? If yes, would you mind pointing towards a short project on single-cell analysis based on public data. I'm trying to build a portfolio to show that I'm genuinely interested in this field. Thanks for your time and help!

[h.mon]

Do you think if it is ok, doable and practical?

Yes, it is, but be sure to get a good background on biology.

[brc]

Thanks for your message. Would you mind sharing some books or other resources for this purpose? I've started with the book "Molecular and Cell Biology For Dummies".

[Ahill]

Welcome, I'd say go for it. Your physics skills in analysis and modeling are very germane, and of course there is a long history of physicists transitioning successfully into biology and bioinformatics. If you don't have previous cell or molecular biology experience - invest the time to learn. Re datasets, the good news is that there is a lot of available data, the challenge is that quality of data and metadata vary. But having said that, you could take a look at other Biostar threads like this one, grab some data, and dive in.

[brc]

Thanks a lot for your encouraging words! Thanks also the link. I would like to place the same request again: Would you mind sharing some books or other resources for molecular and cell biology for beginners?

[Joe]

The best thing to do would be to find the core textbooks for an undergraduate course. For me, it was Cell and Biology.

This is, of course, a massive undertaking, so it may be best to find the particular area you are interested in (e.g. RNAseq) and dive as deep in to the background literature as possible.

[brc]

Thanks for sharing your thoughts. The link is not working. I was wondering if you are talking about this book: https://www.amazon.co.uk/Molecular-Biology-Cell-Bruce-Alberts/dp/0815344643/ref=sr_1_1?keywords=cell+and+biology&qid=1551949669&s=gateway&sr=8-1

[Joe]

Ah sorry about that, not quite sure what happened there.

Yep, that's the one I (and others often) referred to as "Cell" (https://www.ncbi.nlm.nih.gov/books/NBK21054/)

The "Biology" I referred to is Campbell and Reece's: http://www.mypearsonstore.com/bookstore/campbell-biology-0134093410

[brc]

Thanks again for the references.

Answer 1

[Ali Jam]

Machine learning is a wide area which can be applied in different manners. With a background in physics (what branch of physics?) you need to be familiar with biology. Then you can outline a fantastic mutual research project consisting of biology and physics. for instance, protein-protein interaction PREDICTION. you can perform different prediction research in biology through machine learning.