Ask HN: How can a Web Developer get into Bioinformatics? | Hacker News Reader

Ask HN: How can a Web Developer get into Bioinformatics?

70 pointsbioweek40 comments

I've been doing web application development (and general programming) for almost 10 years now. I've also done a lot database design, working with very large databases, and optimizing queries, etc. But I have no biology experience.

I've always wanted to work with new biological technologies, for example figuring out what genes do, or how proteins fold, or anything cool like that where I can increase the world's knowledge.

It seems to me that a good programmer would be useful for these types of jobs. Does anyone have advice for how to enact this career shift?

View on HN

That's exactly what happened to me, but I have to say that it was more of an accident than a planned move.

I was looking for a job where they'd hire people with Python skills and that could provide sponsorship for a H1B. The intersection of these requirements proved to be quite small, but I was fortunate to find a research lab affiliated to Harvard Medical that did just that.

My first project has less to do with bioinformatics and more with secure manipulation of patients' EMR. It's funded by the CDC and all of the code is GPL.

I think the other project I'm getting involved is more according to what you are looking for: Galaxy (http://galaxy.psu.edu/) is a really cool system where you can process and share huge amounts of data. It's also an open source project(http://bitbucket.org/galaxy/galaxy-central/wiki/Home), and Penn State is hiring.

We started using Galaxy a couple of months ago in my company. I'm not involved in the projects using it but I've seen a demo and it seemed like a cool software, although it has its glitches, according to my colleagues :)

Bioinformatics is cool, but it's basically less sexy than you might think. There are various directions to work on, but the kind of work is essentially statistics and pattern matching. If you want to figure out what genes do or how proteins fold in general, go read a textbook. Spending a year on a single gene processing gigabytes of dirty data may not be as sexy as you might think. Or it might be if that's your cup of tea, but as I said, in the end, it's mostly statistics.

Interesting. Pattern matching, eh? I'm pretty good with regular expressions... sign me up!

That's a very poor representation of bioinformatics. There's certainly a lot of data munging and cleaning (like in any field with lots of data), but bioinformatics/computational biology is not about dumb pattern matching or just writing code. Biologists can code just fine (I got a job at Facebook off a PhD in computational biology).

It's all about data mining and machine learning on biological data, to gain real insight about the functioning of an organism, often then doing the experiments to test that insight. I've seen lots of computer scientists and programmers jump in and flounder, because they're under the misconception that all it takes is the application of canned algorithms to a new data set. You need to understand biology (genomics, molecular biology, genetics, structural biology), first and foremost. Then you need to understand a significant amount of statistics, information theory, and computer science.

This isn't meant to discourage you, but you're not going to get far beyond doing more web programming for biology websites if you don't teach yourself some of these areas.

Coming from having done a Bioinformatics PhD and Post-Doc, I reckon getting in touch speculatively with people in Bioinformatics should work. You might get a lot of knock-backs but there's definitely a need for people with IT skills and a genuine interest in Biology. I'd also say that a strong grounding in statistics would be the biggest selling point - probably greater than a knowledge of biology. Good luck, I hope it works out for you.

I'll second the "strong grounding in statistics" point. Most computational bioinformatics ultimately boils down to statistics in some form or another...

A few years ago, I was talking with the director for bioinformatics and such at a reasonably large medical research university. He told me they really don't hire "pure" computer science people any more, because there are increasingly academic programs combining CS with biology, and they much prefer to hire people who are skilled in both arenas.

That said, I agree with you; it ought to be possible for a good programmer to get into the field and be useful. But gaining more experience/knowledge about the biology side of things may be an important ingredient.

What you're describing is the life I left to start my latest co. I worked at a big research university having not had bio since college .. years ago. The key is to find ways you can help, find someone who needs what you have to offer, and work like crazy to get them to like you. Not that I did this, I kinda fell into it, but I watched others do it with much success. That said, as others have mentioned, statistics, statistics, and then when you need break, some more statistics. Learn Python and bioPython, learn NCBI Blast, and WU BLAST inside and out (read the papers), know some bio basics, and prepare to take a pay cut.

Also, if you find a group you're interested in, who's work you like and can get to interview, or at least a meeting / talk with anyone, read every paper to come out of their lab since the dawn of time. These papers can be hard to grok, and down right painful to read. Read them, and read them again. Show interest in their work and they will like you. It's ok to not get it all. It's ok, to not truly understand what they're trying to do. Trust me, some reviewers don't seem to read the entire papers either ...

If you have any further q's, you can derive my email quite easily from my profile. Ping me. Also, where do you live? If you're in the Boston area, we can talk in person. I know a LOT of people in this space and we were always looking for great software engineers who had a thirst for biology.

You can get into bioinformatics without a background in biology, but you'll need some math. Assuming you have basic calculus & some statistics, I'd recommend looking at the Univeristy of Illinois at Chicago's online certificate in bioinformatics (http://proteomics.bioengr.uic.edu/online/). You take an intro course and then choose two more from among computational genomics, molecular modeling and biostatistics. It's a great program. They also have a masters of engineering degree with a bioinformatics specialization, again completely online.

If you're looking for something more immediate, get a hold of a couple books and familiarize yourself with the basics. Mount's Bioinformatics is great, and it's the text for the intro class at UIC. Other good ones are Bioinformatics Algorithms, Python for Bioinformatics, and the O'Reilly books on Perl for Bioinformatics.

Finally, hook up with a bio professor who needs some tech help and volunteer. You'll learn a ton.

You may not have luck finding work in the field without knowledge of biology or chemistry or statistics or maths. There are many graduate programs that focus on retraining software developers as bioinformatics researchers. Also consider cheminformatics.

Please send me an email [1] if you want to learn more. The lab I work in is always looking for eager new students.

[1] http://mailhide.recaptcha.net/d?k=01kUsXu40I09lRKAR_qd4ZFQ==...

I have a friend who is doing his master's thesis in Bioinformatics. If you like running batch jobs that can run days and fail in the middle, then its a great discipline for you! All kidding aside, the local University might be a good place to start. Often working as a grad student you can learn the books as well as get practical knowledge when working with a mentor. If you're near Orlando, FL I can get you in touch with my friend at UCF.

I keep finding interesting things out about central florida--does Tampa Bay have anything interesting in that area?

I'm not familiar with USF's program. You might be able to contact their people and see though.

I am in an undergraduate-level bioinformatics course right now, and can echo what was said about statistics. The biology isn't too tough to pick up, but I am weak in stats and that is my greatest difficulty with this class. That said, I am enjoying it immensely. The book we use is a good overview of the subject - http://www.amazon.com/Bioinformatics-Molecular-Evolution-Pau... -Bioinformatics and Molecular Evolution. Most of the tools that bioinformaticians use are buggy, horribly designed and very unfriendly to the average user. One way to get involved is just improving the usability and stability of the tools. There are plenty of command line tools that you can add a GUI to if that is your thing:

http://bio-bwa.sourceforge.net/

http://abacus.gene.ucl.ac.uk/software/paml.html

http://samtools.sourceforge.net/

You don't need to know too much to contribute. If you are in Utah and interested I could get you some contacts with professors.

I had the exact same desire about a year ago; I was bored with web development and wanted to learn a new discipline (I never went through with my plans though, so I have no real advice for you). My planned route was to get a formal education at the university. I'm sure there are self-taught bioinformaticians out there, but my perception of the field was that it would be very useful to take at-least a few courses to get into it.

Another possible angle is high performance computing as it relates to biological data.

Today's DNA sequencing instruments can kick out tens of millions of data points. These can be mapped back to where they belong on a genome and then from that data, variations can be calculated. Want to find out what mutations a cancer patient might have? This is how.

This would be a perfect place for someone who might be interested in cloud computing, for example.

I work in the field myself, and on my team there is a mix of pure CS people and computational biologists. In my experience there are plenty of opportunities for good developers to get in the door and make huge contributions without necessarily having to have all the comp bio experience.

And, for what it's worth, data visualization is always a problem and the field will always need sharp web guys to develop good tools to view what's going on. This gets especially important now that the volume of data that you're trying to get answers out of is so ridiculously huge.

If you're contacting people about this (and you should), try to do some reading beforehand. In particular this bit:

I've always wanted to work with new biological technologies, for example figuring out what genes do, or how proteins fold, or anything cool like that where I can increase the world's knowledge.

What about Bioinformatics interests you? More specifically: exactly what part of the world's knowledge would you like to increase and why?

I think a good, detailed answer to that question (and decent coding skills) is the most useful thing for you right now.

Maybe this is useful?

http://ocw.mit.edu/OcwWeb/Electrical-Engineering-and-Compute...

And here are links to courses (and therefore course notes) at Stanford:

http://www-helix.stanford.edu/people/altman/bioinformatics.h...

Pick up a copy of "What color is your parachute?" Read up on the advice on informational interviews. Do your best to find some contacts in this field and see if they will agree to an informational interview. (Rule #1: You are NOT networking at this point. There can be no intent to finagle a job this way.)

Someone higher up than me at my company (not a bioinformatics thing) told me that I might qualify for some positions that it doesn't look like I would because I have a strong-ish math background and they basically look for that and do the other training in house because they simply can't find people who already do what they need done. So you might find that some jobs in bio-informatics follow a similar pattern. This information is unlikely to be public knowledge or advertised openly in job descriptions. You will need the inside scoop and that's where an informational interview would come in very handy.

Good luck with this.

That book is absolutely amazing. It's almost a secret weapon for finding work in that it's completely indirect. If you talk to enough people who do the kind of work you want to do, eventually someone will see potential in you and not necessarily offer you a job, but KNOW someone that might. Just talk to everyone that's willing to give you the time, be honest and enthusiastic about what you want to do (and having enough background knowledge to sound competent sure helps), and eventually you'll make the right contact. It's almost certainly not going to be a quick process, but it's an incredibly powerful one.

You don't need courses solely in bioinformatics. You need a solid foundation in molecular biology and genetics. These will give you the requisite background to speak intelligently about the issues. You can take additional courses in (or read independently) about bioinformatics concurrently.

Depending on the project, there is a good deal of statistics, as well as biology knowledge.

As far as getting involved, I would check out current literature in the general area of "Bioinformatics". Also consider searching for publications in Systems Biology, Computational Biology, Functional Genomics. Usually there are good articles coming out of Genome Biology, BMC Bioinformatics, BMC Genomics, as well as Nature, Cell, etc.

I recently began a position as a Bioinformatics Software Engineer, after spending 8 years on Wall St building trading and financial analysis applications, which was somewhat of a switch (though my bachelor's is in Chemistry). Spending time reading the current literature has helped me get involved/engaged.

There are definitely opportunities for someone with your skills to ease into bioinformatics. Having conducted a job search not long ago, they are unfortunately a bit hard to find at times, depending on where you're located. In general, I think you'd have good luck starting with some of the less science-heavy bioinformatics work. For example, I'm at The Center for Biomedical Informatics in The Children's Hospital of Philadelphia Research Institute. Our group spends a lot of time meeting with researchers, understanding what they need, and building (or buying) systems that solve some of those problems. The team I am part of is much more focused on biomedical informatics than pure bioinformatics (though we have people who do "pure" bioinformatics as well). Our setup really makes things interesting. Many of the "big questions" in bioinformatics are real tough problems to crack (such as protein folding). But there is practically an unlimited need for people to have many smaller problems solved (with great impact). Simple things like integrating research data into a database-backed web-app can give a clinician the tools he or she needs to conduct their research more efficiently and uncover new hypotheses. Plus, we get to work on things that directly impact the health of kids. Many times, we're working directly with a physician who is conducting research into a particular disease. They literally will be seeing kids early in the day, and having a meeting with us about UI design later in the day.

We are actually hiring for two positions right now:

Senior Data Integration Analyst: http://www.chop.edu.apply2jobs.com/index.cfm?fuseaction=mExt...

Analyst/Programmer http://www.chop.edu.apply2jobs.com/index.cfm?fuseaction=mExt...

We're always looking for the right kind of people and I maintain that hacker news readers are a good pool to draw from. Don't let the corporate sound of those descriptions fool you, we're a pretty entrepreneurial group. If by any chance you're within easy travel distance from Philly, or even if you just want to chat more about bioinformatics, feel free to email me at italiam [at] email.chop.edu.

There are plenty of approaches that are helpful for programmers in both Web and Bioinformatics domains. Seriously.

Check out Programming Collective Intelligence, and I bet it will give you some ideas.

You would be referring to the book written by Toby Sagaran would you? If it is I second your notion, an excellent book btw. Really got me into algorithms and the like.

There are a lot of bioinformatics positions out there that don't require much biology out of the box. Many universities have bioinformatics departments with lots of programming positions. I worked for the past year as a student programmer helping develop and maintain the data entry website for a set of international clinical trials. My superior was a seasoned Unix/Perl hacker who found a niche coordinating the development of clinical studies websites.

Some of these jobs look like a good place to start in the field.. http://www.bioinformatics.org/jobs/ The best way to get in is to find what these companies are looking for and start honing your skills in that direction.

Might be worth checking out these guys - http://diybio.org/

See if theres a local group and go along and talk to a few people. I'm sure they'll have some stuff you can get your hands dirty on that will at least get you some experience in the field

http://www.tcoffee.org/Courses/Exercises/index.html

http://www.larrygonick.com/html/pub/books/sci3.html

Don't laugh, excellent books!

There are many current opportunities in bioinformatics. The discussion in this thread seems to have some valid commentary. One place to check out might be

http://genome.wustl.edu/careers/

We are hiring. jeldred@genome.wustl.edu

I made a similar plunge several years ago. My best advice is to look for a position in a research group that does both wet-lab experiments and computation. They may be very interested in your web background.

http://arcanius.silverfir.net/blog/clinical-research-experie...

Maybe this guy can help.

Maybe read a textbook on bioinformatics, implement a couple of non-trivial programs/algorithms, and put them on your resume?

Do you need biology knowledge? I have studied in school a little about genetics, Mendel's laws and things like that...

do a search of recently awarded government grants on bioinformatics, particularly small business grants (SBIR). awards are all public, so you should have no problem finding them. contact those folks and just tell them honestly what you are interested in doing (be sure to read their awards first so you have some idea what work they're doing). everyone respects people who are trying to learn new skills, and i think you'll get a lot of warm receptions.

That sounds like a good idea. I'll give that a try.

The results of a search like that would make a great post. Please share what you find.

Do you know how many of these have been awarded in the last six months - especially with the stimulus package? Seriously, this is intractable. Find folks who are doing work that interests you, and then try and find their grants and their papers.

Depends. If he can narrow down the search by finding a specific area of the field he wants to go into it might be more tractable.

I clicked on the 'past awards' link on this page http://www.sbir.gov/

and after lots of signing up and applying for things, I think they 'declined' me to view the past awards because I'm not a government employee. Am I doing something wrong? Is there a better place to look this up?

Reply on news.ycombinator.com