Forum:Recommendations of Linux distros for beginners in bioinformatics

Question

[ropolocan]

Hello Biostars community,

I have trainees who are learning how to use bioinformatics tools for microbial comparative genomics. Although they can do a lot in Geneious, and they are learning a lot from working with it, I would like my trainees to start working on a Linux distro soon. By doing so, I think they can get a lot more out of their learning experience and they can have the opportunity to try more bioinformatics software.

When I was a beginner myself, I got introduced to Linux Mint, and then I tried BioLinux (Ubuntu). However, I have read comments that the distro that might facilitate the most the transtion from Windows to Linux for beginners could be elementaryOS. I have not read any comments about elementaryOS used in bioinformatics. Any comments on working with elementaryOS? Which distro would you recommend for beginners?

[GenoMax]

Are the participants expected to install the distro themselves? On bare metal or as a VM in windows?

[ropolocan]

They will be using laptops that I will set up with the distro.

Answer 1

[GenoMax]

Then why not stick with Ubuntu since that seems to be most popular/has the largest support community.

[steve]

I second this.

Not sure what the OP's setup for class is, but when I started the professor had a single Ubuntu compute server for the entire class to use, and the students could access it from any ssh-enabled computer (Windows, Mac, ChromeOS, etc). This had the added benefit that the students needed to learn the Linux command line interface, which is much more useful for bioinformatics than the DE or GUI that might come with a desktop version.

[ropolocan]

Makes sense. Thanks, @genomax.

Answer 2

[Charles Plessy]

For bioinformatics, I recommend Debian. Our Debian Med project packages a large number of bioinformatics tools, and has regular contacts with members of the bioinformatics community with its yearly sprints (2015, 2016, 2017) or the Google Summer of Code, etc. Being our downstream, Ubuntu has the same packages, so if you can not use Debian, Ubuntu is also a great choice, but Debian is the distributions where the things happen. In case one wants to report, suggest, contribute, etc, it is better to use Debian from the start.

[GenoMax]

Is Debian going to be easy to use compared to Ubuntu for novice's?

I am asking the wrong person :)

[Charles Plessy]

Indeed, I have not used Ubuntu for ages, but my current experience with GNOME3, the default desktop in Debian, is that it is very simple to use. And at conferences, I plug my Debian laptop to the VGA and HDMI connector, and it Just Works, as it should with any OS. Lastly, importantly for the novices, Firefox and Thunderbird are back.

[ropolocan]

Great to hear that your Debian laptop just works with HDMI and VGA. Recently switched to Debian (with XFCE desktop environment), and I am really enjoying it. Thanks for your earlier suggestion on Debian Med, @Charles Plessy. It sounds very interesting.

[GenoMax]

Can you add BBMap to the packages?

[Charles Plessy]

There has been at least one attempt to package BBMap, but apparently it is difficult. The issue here is that in Debian we recompile everything from source, to propagate bug fixes trough dependency chains, to demonstrate that the software is concretely Free, to leverage some upstream regression tests that are run at build time, and more recently to test for bitwise-reproducibility of the build process. This unfortunately does not fit very well the usual practices in the Java ecosystem.

Answer 3

[Farbod]

Ubuntu:

The leading operating system for PCs, IoT devices, servers, clouds and the next technology that I do not know its name, yet!

[ropolocan]

Nice image, @Farbod. Thanks!

Answer 4

[anicet.ebou]

Ubuntu.

Because of the community and all the ressources available it will be more easy for your student to get in touch.

When i was on my first project (Transcriptomics to predict toxins peptide-like components in venom gland of a spider), the community help me a lot to master in few days basics on how to use the terminal (bash, perl, Bioperl, python, Biopython) and others utilities like samtools. Also some bioinformatics tools mafft, hmmer and blast used locally to keep confidentiality.

Answer 5

[dago]

Hi personally think that the distro does not really matter. I would choose a well supported one which is mainly compatible with the majority of Bioinfo programs. I like the Devian-based one, like Ubuntu.

Besides Biolinux, I would strongly suggest you to take a look to Bioconda. In that there are plenty of packages and programs pre-isntalled and this will help a lot in the begging, because will avoid a lot of pain for program installation. In my case this was the major waste of time when I started play around with the black screen.

Hope this helps.

[WouterDeCoster]

I believe biolinux is no longer updated/maintained? The latest release is from 2015-01-09.

[dago]

I think so, I am planning to update my system and try Bioconda. It looks really well planned and updated!

[WouterDeCoster]

But Biolinux is an operating system linux flavor and Bioconda is a package manager. They're not something you can compare. But Bioconda is great. I use it all the time on my Ubuntu17.10 and CentOS6.

[dago]

Yes you are right, but practically speaking they both offer many bioinfo packages and programs, offering a solution biologist-friendly for informatic problem potentially time consuming and frustrating.

[ropolocan]

Agreed. Bioconda is one of my favourite solutions for managing bioinformatics software, along with Linuxbrew. I enjoyed working with BioLinux, though I haven't kept up to date. They had a great tutorial for introduction to Linux from which I learned a lot.

Answer 6

[cpad0112]

Any main linux distro is fine (debian based or rpm based or source based). Refer to distrowatch for distribution ratings. Now that much of the bioinformatics analysis is done via frameworks (make, Snakemake, bpipe, clusterflow etc), interpreted languages (python, perl, go, julia, ruby, scala etc), shell scripts (bash, fish etc), IDEs (R/Rstudio, Rodeo for python etc) and web based (galaxy, jupyter etc), distros do not matter, IMO. In addition for installing software (in user land) brew, synaptic, apt and other efficient package installers are available.

If you can afford, AWS services are available with enough bioinformatics AMI images. I would go rather with any (linux) distro that is supported by local IT core.

Answer 7

[Arup Ghosh]

Biolinux in a bootable external drive or VM is great option to get started in bioinformatics. Or to introduce some difficulty you can start with any Debian or RPM based operating systems like Ubuntu or Centos. But you should choose an os with good community support else the Windows to Linux transition may be hard for the participants.