Forum:What Skills Do You Have To Have As A Phd Graduate In Bioinformatics?

Question

[sentausa]

Hi. I don't know if I may ask this here. I'm going to start my 3rd year in PhD study in a country where PhD study usually lasts for 3 years, and I feel that I haven't developed more skills than I expected.

I'm working on comparative genomics of a certain species of bacteria and my tasks include finishing the genome sequencing project (i.e., closing gaps by PCR & Sanger sequencing, which I have to say I don't enjoy, or by dry lab way using CLCgenomics to locally reassemble the reads with a reference genome), doing annotation (the actual functional annotation is done by my co-supervisor with his special pipeline, I do only ORFing and curating the pipeline output), and comparing the genome content to 1 to 3 closely related bacterium (I use OrthoMCL to find orthologous proteins and determine which bacterium doesn't have which protein). The results have been published in several genome announcements of a journal, and I still have to prepare "bigger" papers to explore more on the genomic comparison. However, I'm feeling that I haven't learnt enough.

When I see the requirements they're asking for postdoc positions, I'm terrified. Everyone asks for fluency in some programming languages, which I thought I will acquire while doing PhD (I learnt Perl & Java in my master's, but it was such a long time ago...), but I never need to do any programming up till now (or I've been too stupid to not know the contrary).

Am I being undereducated or is it just my feeling?

UPDATE: I'm happy to tell you that I have found an opportunity to use programming in my research, as I put in this question. I am planning to use Python to solve the problem.

[jivarajivaraj]

when I read your post I disappointed for myself more and more because I'm a 3rd year PhD student now while I even cant do a simple blast...meanwhile my thesis is in bioinformatics field...when I read your experts I asked myself how u can do your thesis?!!!

Answer 1

[Damian Kao]

You will need to learn a language and be fluent at it if you want to be competitive for a bioinformatics job unless you have some very niche or special knowledge. While available software and packages are great, the field is moving so fast that you just can't get way with only using other people's pipeline or software.

My advice is to get started on learning R, python or perl right away. I am on my third year phd also. Even though I have a decent grasp of programming and theory, I still feel extremely unprepared. Especially when reading the stuff other people are doing on blogs/papers.

[Richard Smith]

Agreed. I would look for any steps in your workflow or analysis that can be automated by a python/ perl program or a shell script, even if they're trivial to do manually.

Once you get started I expect you'll see many other instances where you can benefit from writing programs (and can write on your CV that e.g. you automated your analysis with python :)

[sentausa]

Thanks for your reply. I understand what you feel. I even have no ideas about what they're asking here, but I'd like to learn.

Answer 2

[Gabriel R.]

The most important skill of all is learning how to use the bash terminal effectively along with normal commands like awk/sed/cut/head/tail etc.

Answer 3

[Leonor Palmeira]

I agree with Dk, and would advise you to talk this over with your supervisor and/or co-supervisor. Bioinformatics is definitely much more than just running a couple of tools and interpreting the output. Would it be possible for your co-supervisor to let you work with the functional annotation pipeline? Maybe also tweak it and test different things? Maybe there are some aspects of your work that you can use to train yourself in Python or R?

This being said, I see you also have acquired wet lab skills, this is a big plus! Maybe you should decide whether you want to focus on bioinformatics (and learn the techincal skills that go with it) or on molecular biology (and deepen your knowledge there).

And, you can have a look at this other question which is somewhat similar to yours in my opinion.

[sentausa]

I'm sure that I won't focus on wetlab (in fact, one reason I decided to study bioinformatics is that I just can't stand handling chemicals and wet things :p).

About my opportunity to train myself in Python or R, I'm wondering if there are comparative genomics techniques that might need programming, since I will still need to make further comparisons on my bacterial genomes.

And thank you for directing me to the other question; your comment about reading materials & methods section of papers is quite interesting.

Answer 4

[Alex Paciorkowski]

One way to look at this that might be helpful is to ask yourself where you want to be in 5 years and in 10 years. This may help you understand what you really love to do -- and should then help you know what tools you will need to get there. Do you want to be a PI running your own combined wet-lab/dry-lab research program? Do you want to be a deeply-knowledgeable niche collaborator, who is able to secure funding and publish papers because you and only you know your corner of the field the best? Or do you want to run a bioinformatics core at a research institution, perhaps being a jack-of-all trades, the go-to person for multiple types of informatics tasks? Each of these is an awesome career path for the right person -- but they involve different skill sets. Find colleagues who have have been successful in a variety of bioinformatics career paths -- and see what skill sets they have that allow them to be successful. Ask them when and how they learned what. And don't worry if you feel you are "starting too late." I think we've all felt that way about something.

Either way, fluency in several programming languages will serve you well along any of these career paths. I agree with the prior answers -- find a way to dust off your knowledge of Perl and Python that can help you analyze data and get papers published. Talk to your supervisors about ways to do this. Perhaps there is a corner of your current thesis work that would be amenable to development of a new tool written in either of these languages. The best way to learn to use these languages is to use them to solve some problem that comes up in your own work that interests you (IMO). The same thing with R. Good luck.

[sentausa]

Thank you very much for the answer. I wonder if you could share here on what it took for you to get to where you are now in bioinformatics. Or perhaps Leonor and Dk or others could share their own experiences here?

Answer 5

[Dan D]

It seems like your main point of concern is your perceived lack of applicable programming knowledge. But you have background, and that's great! Here's my recommendation: decide which language you want to use, and then visit rosalind.info. It's a collection of computational problems in the bioinformatics realm. You won't know the answer immediately to most of the problems, and that's the whole point. It will give you hints and give you starting-off points. The problems start off very easy and gradually get more difficult, building on previous ones and categorizing them. You fill in the gaps in your programming knowledge with experimentation and researching the language you're working with. In many ways it's faster and more effective than getting a generic book on a language.

By the time you do the first ten problems, you'll have much better grasp of how to use a language's data structures in a bioinformatics context. Then you can honestly tell prospective employers that you have competency in a language. You can use the knowledge you gain and code you write from these problems in real-life situations too. I know I have.

EDIT: Let me go ahead and add that, based on my experience, you aren't going to learn much practical coding knowledge as part of your curriculum. Programming is about 80% of my job and I learned all of it on my own. If I can, you can too. It just takes a lot of patience at first. It becomes worth it very quickly as you will greatly increase your problem-solving skills in general after wrestling with code for just a few months!

Answer 6

[Naresh]

To start with a PhD in Bioinformatics - one has to have a zeal to work hard and learn things. For any PhD. We don't need to have skills, but the interest to work and find out new things. that is Research..