Question

Should a Bioinformatics PhD student compulsorily conduct wet-lab techniques for PhD?

4

Entering edit mode

8.3 years ago

Govi988 ▴ 80

Hi all,

I am working as a bioinformatician in one of the genomics lab from past one and half years. Although, I theoritically know biology and understand the experiments, I do not have experience working in wet-lab prior to this or at present. I am now thinking to join another lab, where I got a PhD position. Although I will be a student, the lab would look me more likely as a research assistant rather than a PhD student because the data generated from the lab is huge and a bioinformatician is very much needed for them to analyse the data.

With these circumstances, I might still end up getting a PhD, but I would have completely depended more on other people data and wet-lab ideas rather than my own. I am sure that I will still get good journals and become much better analyst of the genomic data bioinformatically, but I would like to know how many labs would want such kind of as post-docs or RFs, who does not have practical knowledge of the wet-lab techniques after finishing PhD? Should a bioinformatics PhD holder compulsorily do wet-lab? May I collect all your valuable ideas. Thanks very much in advance.

next-gen phd genomics • 3.6k views

ADD COMMENT • link updated 8.3 years ago by DG 7.3k • written 8.3 years ago by Govi988 ▴ 80

score 4 · Answer 1 · 2016-01-21

My answer would be that if it is something you want to do, then by all means gain that experience. If the wet lab portion doesn't appeal to you that much, then don't. As Emily said in her answer, it is very important to understand the biology that is happening and the science behind the various experiments, and why they are conducted as they are, and what effect that has on the data. It isn't always necessary to do wet lab work yourself though to have that understanding. At some point both the bioinformatician and the wet lab biologist are dealing with theory and inference about what is happening at the molecular level. Neither one has a privileged understanding of that. You can do experiments like a robot just as easily as you can analyze data like a robot. Wet lab versus computational doesn't give you more or less understanding (necessarily) of what is happening. It just gives you more experience about conducting the type of experiments you conduct in order to study the biology.

As a side note though, you want to be VERY clear going into a PhD program about the sorts of things you'll be doing as a student in that lab and what the training expectations are. Are there other bioinformaticians in the group or does the PI have bioinformatics experience themselves? If not then at the very least I think you would want to look at a co-supervisory position with another mentor. I did my Post-Doc as the embedded solo bioinformatician of a group, which was fine for me as I already had very strong bioinformatics connections for additional mentoring and just needed interesting projects to work on. It led to my current faculty position because I became very valued doing what I do. However, as a PhD student I wouldn't recommend it because it will be hard to build a large skill set, form connections with other bioinformaticians, and most of all, develop as an independent scientist. You really don't want to be just doing computational analyses and running pipelines for other people's projects, you also need to be doing your own projects.

Ram · Answer 2 · 2016-01-21

This is an excellent topic of discussion. I am a professor who mentors PhD students. My opinion is yes. It is important to have hands-on experience with the nuances of experimental work to understand sources of variance...both biological ("real") and those caused by experimental/user error. Even if the experiments are simple...such as drug-treated vs. non-treated cells grown in culture, wet-bench research helps you to better discern when certain types of data analyses are appropriate.

One example that comes to mind is the current problem of reference genomes for cancer research. Less than five of the bioinformaticians I have ever interacted with (out of maybe 30 or so) actually understood the essential underlying problem of using a wild type reference genome to align ChIP-seq reads from established cancer lines. Recently I had some ChIPseq done on U2OS cells and the experts at the core facility could not offer me an explanation of how the alignment could be run to compensate for mismatches. Maybe they did nothing about it at all.

Anyways, you may not have to grow actual cells to grasp the general idea that cancer genomes are mutated, and can result in gaps in alignments, but having hands-on exposure to the biological side of things and conserving regularly with wet bench biologists can help to broaden your horizons in other important ways.

I hope this helps.

score 3 · Answer 3 · 2016-01-21

I don't know if doing is necessary, but I certainly think that it's necessary for a bioinformatician to understand the biological techniques involved. For example, a bioinformatician who has always previously worked with RNASeq data can theoretically start working on ChIPSeq data without any problems, if they know the software to use. But if the bioinformatician does not understand the wet-lab technique in ChIPSeq, specifically the random shearing of the DNA, they will struggle to understand why their protein-binding regions have such fuzzy edges.

You might find the best way to gain this understanding is to do some wet-lab work. You might find that just attending lots of talks by wet-labbers, and properly discussing any data that comes up along the way is enough.

Ram · Answer 4 · 2016-01-21

To add to what everyone else has said:

I think another advantage of learning the "wet side" is that it increases your overall understanding of experimental biology. My experience has been that it gives you a better idea of how different approaches complement each other. How wet and dry methods together can build a more complete picture together. I think it will allow you to further your own work by building more complete stories as well as interact more effectively with collaborators. For someone who wants to pursue research as a career this is incredibly useful.

It will also help you understand what the limitations are for the wet front end of experiments. Trust me, it's easy to tell someone you want half a dozen replicates for four conditions along 5 time points with a randomized plate order and to the minute timing data when you're not the one running the experiment. It's another thing to actually have to run those types of experiments.

Probably what has helped more than anything is the grounding it has given me with respect to the capabilities of bioinformatic and computational approaches. Not so much learning what wet methods can do, but learning what dry methods can't do.