Phd Candidate Interview Questions For A Mixed Lab/Bioinformatics Project

Question

[dan.brewer]

I am the boinformatician on an interview panel to recruit PhD candidates for a mixed project in cancer biology where the PhD candidate will be performing lab-based experiments and applied bioinformatics on mainly NGS data (I have no control over this strange mix). For the bioinformatics part it should be only the use of existing tools and performing some statistics. I would like to ensure that the candidate doesn't struggle with the bioinformatics side of things. Has anyone got any idea of the type of questions I should ask?

I think there is three main areas to cover, programming/computing, maths/statistics and bioinformatics.

[Istvan Albert]

I think this is a bit one of those pie-in-the-sky expectations. Who wouldn't make a strong effort to recruit students like that? In reality it is more about getting lucky rather than recruiting effort - unless you are in one of the top ten universities that get an unusual number of applicants (and high acceptance rates).

The only realistic approach that I see is to provide that training on site within your organization where you could educate any hard working and smart student to get these skills within a reasonable amount of time.

[Medhat]

what a bout giving a simple real task from your project?

[Damian Kao]

PhD candidate as in interviewing for potential new PhD students to work on a project?

Answer 1

[Emily]

Given the wet-lab side to this project, I would guess that you're looking at people straight out of their undergrads or masters, studying life sciences. It's quite likely that the applicants won't have any prior experience in bioinformatics. If you're lucky, you might get some recreational programmers but I wouldn't count on it. From this, I would be looking for ability to learn over knowledge. This means people with mathematical skills, basic IT competency, logical reasoning skills and an enquiring mind.

Answer 2

[Arun]

I basically agree with Emily's answer. In addition, my recommendation would be to first ask what he/she has done and take it from there. For example, if someone has not worked at all on an NGS project, it doesn't make sense to ask them to explain the workflows (and there are/were quite some colleagues who simply were not working / their curriculum did not cover it).

I'd start with their programming skills (ex: python, R, perl etc..). And I'd ask for what packages they've used / found useful in the context of bioinformatics. If necessary, I "may" ask them to write a small snippet doing a very basic task. Again, there's no point in asking packages in python if he/she has used R/perl.

I'd be testing if what they claim they know is true or not "more" than trying to be ambitious in asking if they know every one of the gazillion pipelines that are around. In essence, if they know very well what they've done, it gives a good idea of their integrity and one could find out if they're good at adapting to new environments and if they're motivated.

But that's my view :). Hope that helps a bit.

Answer 3

[Pierre Lindenbaum]

• can you explain the meaning of each SAM flags ?

• describe a simple NGS workflow from the FASTQs to the VCF

see also:

What is your favorite question to ask when interviewing potential Bioinformaticians?

[Ben]

Is the correct answer to the first one, "I'd probably just check the manual"?

[Pierre Lindenbaum]

the manual says "each segment properly aligned according to the aligner" , " PCR or optical duplicate" : what does it mean ? :-)

Answer 4

[k.nirmalraman]

• Explain your bioinformatic approach to a given biological question (certainly something that you do in your group) and expand it with further questions based on the candidate's answer! This will help you get an understanding about candidate's fitness