Forum:Some tricks from the experts

Question

[dago]

I am not a pure bioinformatician. I self learned Unix and R and I must say I have been quite happy with the things I achieved.

However, I realized that I would need some more hands on a programming language. I choose Python, as it seems to be easier to learn for new comers like me.

I started looking online tutorials and wikis, and I found them really useful. However, at some point I start to think that maybe I am just wasting my time in learning basic programming aspects, which I will not really need in the real life when dealing with my data (mainly sequences). I think that this might be a common problem that biologist approaching a programming languages have. Therefore, I was wondering if you have any suggestion or advices concerning which aspects would be more useful to learn for a biologist that starts to learn a new programming language (i.e. Python, Perl).

Thanks for sharing

Answer 1

[WouterDeCoster]

My suggestion would be to lock yourself up in your room and get started with solving questions on http://rosalind.info/ You get (increasingly difficult) tasks regarding bioinformatical questions which you can solve with any language you want. Not only will you get better at programming, but in addition you will get an understanding of how famous algorithms work, e.g. BLAST, by writing things like that yourself.

[dago]

that is a really good suggestion. I knew Rosalind, but I did not spend so much time on it. One of the "problem" I faced often is to work/edit external files. Any suggestion how to improve these skills?

[WouterDeCoster]

It's not clear to me what you mean, do you face problems editing an external file using your script?

[dago]

Yes. Like, load a file and apply specific action, like keep just some lines rather than other and so on.

[WouterDeCoster]

So in R you would read the file in memory using the read.table() function, make some changes and save an object back to the disk using write.table(). In python you would open a file, either in memory or iterating over it, and writing modified lines back to disk. What exactly is the problem?

Answer 2

[Vivek]

A couple of minor tips if you haven't already adopted them:

1. Learning and practicing version control is important albeit often ignored but pays good dividends in the end. Set up Git on your Desktop if you haven't already done it and keep your repositories synchronized. Good organization and proper documentation is a lifesaver if you are writing scripts that you might use for different projects at a later date or collaborate with others in your lab. Even if you are just solving Rosalind problems, putting up code on Git and linking your repository on your resume helps in job interviews where people want to see your code samples before even deciding to call you for an interview.

2. If you write R scripts that you might reuse or distribute to others, get familiar with dplyr and write code in the dplyr way. I found that it instantly makes my Rscripts far more readable for others and even myself when I revisit them a few months down the line.