Forum:Bioinformatics - how to start?

Question

[Abigail]

I'm a student of Molecular Biology. I would like to start studying bioinformatics. Can you tell me what is the best way to start learning, what is necessary or worth knowing? Which areas of bioinformatics are most in demand? What was your career path like? I'm from Poland and there is little information about bioinformatics, so my knowledge is limited.

Please, write whatever you think is important.

Answer 1

[GenoMax]

First thing to start with is linux/UNIX. If you are not familiar with the command line then spend some time on getting this basic skill mastered. There are many resources for learning this but one focused towards biologists is http://korflab.ucdavis.edu/unix_and_perl/current.html#part1.

Learn Python and R basics so you can have languages most often used in bioinformatics under your belt. Catching up with basic statistics knowledge while you are doing all of this is critical. If you are simply interested in learning bioinformatics to supplement your molecular biology skills you would be well situated after you have done these things.

Please do not take the following in the wrong way --> At this point there are plenty of bioinformatics degree granting programs available around the world so unless you are thinking of enrolling in one of these programs you may be at a disadvantage in terms of finding job opportunities (unless you can demonstrate your skills, most visibly via publications, software you may write). So consider this carefully before trying to think of switching to bioinformatics as a source of future job opportunities.

We wish you the Best as you contemplate starting your own journey with Bioinformatics.

[colindaven]

Yes - install Linux locally and force yourself to use it every day (command line).

• hang out on github
• learn bioconda, mamba and how to install software with these
• start programming, eg python, R.
• later - learn a workflow language like nextflow/snakemake
• learn to install and use nf-core pipelines for standard analyses, eg rna-seq

It's lifelong learning, so be prepared to keep it up.

Answer 2

[Mensur Dlakic]

Poland has excellent bioinformatics researchers and facilities in Poznan and Warsaw.

• https://www.iimcb.gov.pl/en/research/1-laboratory-of-bioinformatics-and-protein-engineering
• https://rekrutacja.amu.edu.pl/en/studies-cataloge/bioinformatics,29
• https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1003048

You may want to start by getting in touch with them, and maybe doing a summer internship.

Barring that, there are so many ways to start. While learning Linux is a must for any kind of large-scale bioinformatics, it is not absolutely necessary for all bioinformatics tasks. You may want to start with some kind of class to help with that, or try one of lecture-type Linux tutorials:

• http://www.doc.ic.ac.uk/~wjk/UnixIntro/
• http://tldp.org/LDP/intro-linux/html/
• http://www.ee.surrey.ac.uk/Teaching/Unix/

I think it is easiest to go with your actual research interest and learn how bioinformatics can help, because that will give you double motivation. Not only would you be learning the new field that is of interest, but it could potentially help your current research. Back in the day I was studying unusual proteins belonging to the AP-1 protein family that were different from most other members. Did a bunch of experiments and couldn't figure out why the difference. Eventually, I learned how to predict their structure, and that expanded the original hypothesis and put us on the right track. These days we don't do anything in the lab without having some kind of strong bioinformatics prediction as our starting hypothesis.

Answer 3

[Rajendra KC]

I feel you should take up a Bioinformatics project, that includes some simple and general bioinformatics pipeline like RNA-seq pipeline for a starter. Meanwhile, you could start playing around with Linux and programming language, preferably R. I personally find tidyverse packages of R really helpful in data processing. I feel making good plots is super important in bioinformatics that help you clearly communicate your analysis. You may want to go through some bioinformatics paper to get an idea of various types of cool plots.

Finally, get yourself connected to few online bioinformatics forum, where you will get answer to almost every problem you will be facing. This will also update you about the latest happenings in the field since bioinformatics is an ever-changing field- the tools we used few years back are almost obsolete now, and new tools are being developed everyday.

[Mensur Dlakic]

Meanwhile, you could start playing around with Linux and programming language, preferably R. I personally find tidyverse packages of R really helpful in data processing. I feel making good plots is super important in bioinformatics that help you clearly communicate your analysis.

I know you mean well and I am not trying to pick on you, but this advice is far from universal and mostly reflects your own experience. There is nothing about R that makes it a prefered bioinformatics language to Python, for example. There are probably better statistics and visualization packages in R than in python (and probably some others), but one has to get the data first before making a plot. I don't think it is a stretch to say that python has a larger following and broader applicability in all scientific computing than R, which includes bioinformatics. That still wouldn't make me say that python is preferable to R, but the opposite is not the case either.