Background Knowledge For Cs Person Going Into Bioinformatics

Question

[Persony]

Just some background...

Over the past school year (my sophomore year) I worked part-time as a bioinformatician, at least that was the job title, at the university's life science institute. In reality, I was just a programmer, translating requirements from the PI (I guess that's the right word) into tools he'd find useful. By the end of the year we kind of had a cool toolchain, along with one bigger project that involved me implementing a unique approach to sequence assembly. I'll be returning to this job once the school year rolls back around (could have worked there during summer but...see below).

During this time, I don't feel like I got a very strong background in the biological aspects of what I was doing. The work came in spurts, so the stuff I'd learned from the last project I'd usually forgotten by the time the next idea rolled around. In addition, I worked a lot with the PI to turn his high-level problems into problems that could be solved programmatically, so I wasn't really required to learn the bio aspects of the problems that much. Effectively, I just saw everything as fancy parsing. A lot of the bio stuff I was exposed to went way over my head, and while I was encouraged to learn more about the bio aspects of it ("there's a big future in bioinformatics!") I never did that much besides watching the KhanAcademy bio videos, since I was more interested in my side projects outside work and school.

Currently, I'm interning at a gigantonormous tech company, and don't like it very much. It's considered a "dream internship", but honestly the work I do doesn't provide nearly the same sense of fulfillment as knowing that the biologists used some of my tools to make new discoveries.

With that out of the way, here's my question...

I want to get more serious about bioinformatics. I'm largely a learn by doing kind of person, though I enjoy reading textbooks and papers as much as the next guy (I just tend to forget it if I don't apply it). I know I need a stronger background in biology, probably wetlab stuff (at least an understanding of what's going on), and particularly statistics. The latter is part of the core curriculum, but I'm largely a self-learning kind of person and don't really like the classroom environment.

So, what resources do you recommend I look at to gain a better understanding of these topics, appropriate to the scope of problems I'm likely to deal with? I've seen a ton of recommendations, but have no idea where to start. In addition, what kind of opportunities are there in terms of side projects? What would you like to see? I know web-based stuff has been pushed as the next big thing, but again I don't feel that I have enough background knowledge to figure out what biologists or bioinformaticians would need in a web interface so pretty dry on decent ideas in that area. There's also open source stuff.

If more information is required, let me know. Thanks.

[Chris Evelo]

It might be a good idea to search this site for the "books" tag. There were a number of questions where people asked about books for more or less the same reasons you did.

[Persony]

Thanks. A lot of books were mentioned but good starting points seem to be what drio mentioned and Molecular Biology of the Cell.

[Neilfws]

A lot of bioinformatics really is just "fancy parsing". Some of it isn't even very fancy :-)

[Aleksandr Levchuk]

Just curious, what is the tech company that's your interning?

[Persony]

Microsoft. Made this community wiki since tons of great answers and none are really the "right" one.

[Casey Bergman]

If none of the answer are "right" then perhaps you should consider re-wording your question to provide more specific focus to what you are looking for.

[User 2448]

I didn't mean your answers suck or anything. I mean that the question is subjective, so I can't say one answer is more right than the others. No point in trying to make it less subjective, that is the nature of the question.

Answer 1

[Casey Bergman]

Immerse yourself in evolutionary biology, and in particular molecular evolution and population genetics. These topics are one of the areas where quantitative & computational science interface directly with biology and where you will learn the most about relevant topics (SNPs, GWAS, personal genomics, population genomics, metagenomics, etc.) that can be applied in the post-next-generation-genomics world. Also, to paraphrase Dobzhansky, nothing in bioinformatics makes sense except in the light of evolution.

Answer 2

[Drio]

When you have a clear objective you assimilate and absorb better.

Working in your assembly project would be a great option for you. Sequence assembly (particularly for mammal size genomes) presents some interesting challenges and is currently an unsolved problem. Is there any chance you can keep working on that without having to wait for next year (Perhaps in your free time)?

Another great option, as you mention, would be to join an open source project. It will force you to read papers and books to understand what they are trying to solve. Once you get it, pick a new feature and implement it or do some bug fixing.

I find An Introduction to Bioinformatics Algorithm particularly useful. It covers very well different bioinformatics topics and it does it from the perspective of a computer scientist.

Good luck.

[Persony]

One of my coworkers mentioned that book to me as well, will definitely check it out. Thanks.

Answer 3

[Jorge Amigo]

I guess it all depends on what you would your future expectations, and the timings you are considering:

• if you are willing to be part of a bioinformatics group or to follow your leader, finding the appropriate readings should be your goal. applied bioinformatics books would give you an idea of the typical paradigms of the field, plus it will let you get a little knowledge on the biology involved in the process.

• if you would consider doing your own research, collaborating with a biological group or being part of it (I hope you see the difference between collaborating and following), the best option I could recommend is the one I did myself, which is investing some time in proper academic learning (an MSc for instance). sure this is the most time and money consuming option, but it would open more doors for you if you ever need to move from one job to another.

Answer 4

[julieN]

A couple of suggestions about the bio part. First, you have two more years of school left--maybe there's some room for a few bio/biochem courses; they all include laboratory time. Second, it sounds like your school year lab is what gets you interested. Why not save some of your earnings from this summer's internship and apply for a (much lower paying) wetlab internship in your PI's lab next summer?

On the statistics side, I've always meant to (but never have) get the best possible statistics book I could find and try to code some of the problems. It would give you an opportunity to apply your computer skills to your own statistical problem.

As far as what a real biologist needs in terms of interface, you have the ideal test case in your PI. Try out your ideas (random thoughts, mockups, sophisticated prototypes, depending on what he's like) on him and the other people in the lab who use your software.