How hard is it for a computer science MS student (with a little knowledge in biology) to research problems related to bioinformatics? How much is bioinformatics related to graph algorithms and data structures?
There is a "bio" in bioinformatics for very good reason. In my opinion, bioinformatics that is not focused on helping biologists understand their data and advance biology is worthless. That said, as others have mentioned, there is scope for work on things like algorithm development and statistical methods, but it helps if there's a nice biological problem to which you can apply your findings.
As for "how hard" it is to switch from CS - well, that depends on many factors: the work environment (e.g. how much exposure to biologists), the enthusiasm and motivation of the individual. As a biologist who has learned "the computer science I need to know" as I went along, I'm sure that the reverse is true; a computer scientist can pick up "the biology they need to know" to understand interesting problems.
My hope is that in the future, there'll be no need to debate this question, because science will be taught in an interdisciplinary fashion from the outset. We will just look at different topics as "bits and pieces of things we need to know to solve problems", instead of separate branches of knowledge. I may be optimistic :-)
I will copy my answer from Stack Overflow here:
There's plenty of computer scientists who start a career in bioinformatics knowing little biology. We run a Masters program in bioinformatics, and probably half our intake are computer science graduates.
Taking a computational approach to tackling biological problems is going to require you to have some appreciation of the underlying biology though. If you don't understand the science underpinning the graph, then the greatest algorithm in the world isn't going to help you interpret it any better.
The comp sci graduates on our course spend the majority of their first semester being taught biology.
It depends very much on which end of the spectrum of bioinformatics you are interested in. Some researchers are more interested in the algorithms themselves whereas others, like I, are much more interested in learning about biology by using computational methods. Unsurprisingly, a computer science background is great for the former, whereas it could be challenging to learn enough about biology to be able to do the latter.
The most promising areas to move into with a computer science background at the moment is probably analysis of the massive amounts of DNA sequences produced by new sequencing technologies. In particular, assembly of individual reads to larger contigs is a big problem that draws heavily on graph theory.
The biggest challenge to entering bioinformatics from a computer science background is, in my view, not to be frustrated about the problems begin very complex and at the same time very poorly defined. I think it is fair to say that it requires a change of mindset.
It's certainly possible to jump from CS to bioinformatics, but the more you are interested in biology, and the more you know about it, the better you will fare. IMHO the most exiting results of biofinformatics/computational biology are the biological predictions, not the improved algorithms. Nonetheless, there is lots of room for improving algorithms, even specifically using graphs etc.: e.g. see the to-be-solved problems of terabase metagenomics.
There are three legs to bioinformatics: biology, computer science, and statistics. It's my experience that the biologists can generally get the biological story across pretty well, even if the bioinformaticist doesn't have much of a biology background. This isn't because biology is easy, but because we tend to be looking at very isolated phenomena, and because the bioinformaticist doesn't generally need to know all the deep background needed to make the wet-lab experiments work. On the other hand both the biologist and the bioinformaticist may be relative lightweights in statistics. I'd focus on increasing your knowledge of statistics. (Though obviously knowing more biology is better than knowing less biology).
I would recommend teaming up with practicing life scientists that need bioinformatics approaches. Helping them solve some of their problems is the best way to learn about the applied side of bioinformatics.
As for the answer for your questions: bioinformatics is a very broad field, there are problems that map to traditional CS curriculum, but there are numerous problems that are just simply finding an efficient way to solve a simple problem.
In my experience most problems can be split into a series of very simple steps. The complexity arises from having to make the right decision at each of these steps.
Personally I think it's easier for computer scientists to get into bioinformatics than for bench biologists to get into bioinformatics, especially if it is being done right. That said, it's not trivial and any computer scientist will have to spend some time learning the core tenets of molecular biology. Also, it depends on what your CS focus was. If you come from a data structures, machine learning, etc background, IMO the transition will be a lot easier.
FWIW, there was a related discussion on Quora on this recently
I think all the answers above have great points. I think that it's very important who you work with. Bioinformatics is a very broad field so you can not know everything, but you can work with people who know the biology and you can be the computational expert. That sort of role seems to be coming more common.
In my experience with that scenario, I learn more about the biology and the biologist learns more about the computation but the collaboration is fundamental. I think there are a lot of opportunities to just take some data and run (see the short read archive or 1000genomes for enough data to keep you busy for a while). But even with those, you'll need some biological insight to develop some angle you want to explore.
I can add that somebody with good knowledge of informatics can help to develop user-friendly tools (available on web-servers etc.).
In other words if a research group has well-defined ideas: this program should enable users to do this and this, the expert in informatics can work on this.
I think bioinformatics are much closer to computer science than to biology, that said i mean it is much harder for a biologist to get heavily involved in the field in comparison to a computer science. Most pioneers in the field of bioinformatics come from a purely technological background (mathematics, data science, IT) , very few biologists end up to become real pioneers in this field because it is so heavily data oriented.... A very small background/knowledge of very basic high school biology is more than enough for doing bioinformatics.
PS. Also if you take a look at the job market the employers are looking mostly for candidates with strong computer science background , and much much less with a biology background.
I think this depends on your definition of a bioinformatician. I have always been of the opinion that there are 2 types of "Bioinformatician"
Type I: probably (though not exclusively by any means) from a CS, Maths or Engineering type background. The likes of your Heng Li's etc, who are making hardcore algorithms and software tools with less emphasis on the specific biological question it answers
Type II: your "analysts" and data scientists. The people that use the tools Type I informaticians develop to answer actual biological questions. Probably from an experimental science background, with far more than a high school level of biology knowledge.
Neither can exist without the other.
I firmly disagree with your statement in your PS though. Among the most employable people right now are biologists that have learned to write some half-decent code, for numerous reasons:
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
@neilfws I agree that science should be taught in an interdisciplinary fashion. I am currently taking a Bachelor in computing science specializing in bioinformatics yet except for two bioinformatics classes the computer and biology classes are kept separate.