Fog Creek Software
Discussion Board




New comp sci grauates and bioinformatics?

I am almost thru with college and don't see many opportunities in the traditional software industry.

Do you see bioinformatics is an interesting area for long-haul. They seem to have need for computer science people and relatively nascent?

Newgrad
Monday, February 18, 2002

O'Reilly are getting stuck into Bioinformatics with a new book http://www.perl.com/pub/a/2002/01/02/bioinf.html and a conference http://www.oreillynet.com/cs/weblog/view/wlg/1096

Matthew Lock
Monday, February 18, 2002

i've heard it said that it's easier to go from biology to computers than vice versa since the biology req. is pretty important.  my ex-college roomate had a pretty easy time getting a job at cold harbor springs lab in long island since he had a biology degree and had done work in perl-and this was a guy that failed biochemistry twice!  he turned town a 65 K offer at some firm in salt lake 'cuz of cold harbor's prestige.  of course, this was feb. of 2001-the labor market has cooled off some.

razib khan
Monday, February 18, 2002


I dunno about Bio.  It sounds to me a LOT like "The Mobile Internet", "Wireless Text Messaging", and "The Dot Coms" IE - I whole lot of hype will little delivered.

But I could be wrong.  If you're worried about finding work after graduation, my advice would be to pick up a minor in business, get a high GPA, a good internship, and be willing to more ANYWHERE in the USA.  If you went to a nationally-known college, you should be in decent shape.

regards,

Matthew Heusser
Tuesday, February 19, 2002

A while back, I was at a dinner party, talking to a doctor.  I asked her about some of the problems currently facing medical science, and you might say I got more than I bargained for.  It must have been fifteen minutes of her talking non-stop, enumerating not just general stuff, but specific examples of illnesses to be analyzed, genome sequences to be linked to phenotypes, strains of virii to isolate and identify, etc.  I think I lost five pounds just listening to her.  This wasn't a bad thing, either.  I'm no biologist, but I knew just enough to barely understand some of the problems she was describing.  Fascinating stuff.

The point is, bioinformatics isn't all hype.  There are some real problems to solve there, with tangible benefits (a thousand people cured here, 10000 cured there).  They're apparently facing enough data to make NYSE look like Myrna's Flower Shop, and need a lot of help just to organize it all, sometimes to figure out what a gene does, and sometimes just so a biologist will know what to study!

Paul Brinkley
Tuesday, February 19, 2002

Many programmers can identify a field or domain in which they have experience (for some it might be software tool-making; for others, financial systems; for me it's telecommunications); having some domain-specific knowledge is good (domain experts aren't necessarily programmers, and don't necessarily want to teach you everything from scratch). I'd guess bioinformatics is as good a field as any. It's certainly not true anymore that there's nothing new of interest in biology. Also, it helps that the domain is in an area which people find useful (i.e. there's a continuing demand).

Christopher Wells
Tuesday, February 19, 2002

Just because a technology is in its infancy, doesn't mean that the people working on it are.

I think the most important question is: Apart from the possibility of finding a job in bioinformatics, do you have any interest in it at all?

Lou
Tuesday, February 19, 2002

the thing with bioinformatics is that a lot of the jobs in the pharmaceutical industry-which is more insulated from wild swings than othe parts of the software industry because of lllloooonnnngggg research development arcs.

razib khan
Tuesday, February 19, 2002

I've looked into BioInformatics, as a total outsider, with an inside scoop. A relative of mine was the chief grant proposal approver for the human genome project for 8 years. He's now heading a bioinformatics research lab at a major pharmaceutical.

I , like you, thought the software industry looked pretty dull and going nowhere. And no matter what you all want to believe, in general, this is true. ;-) So I thought I should probably gain some expertise in another problem domain. Bioinformatics looked like the easiest thing to sleaze into.

So, I talked a bit with my relative, and read all the books, and interviewed at some bioinformatics companies, and even talked to the bioinformatics departments at three grad schools. Here's what I found:

1. Bioinformatics is REALLY REALLY BORING. Unless you are into both advanced SQL programming AND genetics , bioinformatics is not going to be for you. I'm not saying that the results of bioinformatics research don't make the world a better, more interesting place. I'm just saying the day-to-day drudge of what bioinformatics research entails is horrendously dull. YMMV

2. The hot bioinformatics guys don't get paid the most money. The Oracle DBA and Sys Admin jobs for pharmas and bioinformatics start ups often pay more money than the PhD head of research job. The admin jobs certainly pay more than the grunt programmer jobs.

This might sound unbelievable, but when you think about it,  most of their programmers are going to be biochem majors who learned perl. so they are going to be real happy making $65K a year re-implementing the BLAST algorithm because previously they were making $12/hr doing PCR gels in some basement lab.

The programmers are replaceable, but the data is not.


3. lots of PERL programming. no value judgements on PERL, just a statement. My relative said that on two occasions they hired some hot java dudes to "port and refactor" the perl piles the scientists cooked up, and each time they ended up scrapping the java and sticking with their old PERL systems.

4. It is the next dot-com disaster job. Hundreds of startups with no business plan are getting VC and popping up in the typical tech startup areas (boston, silicon valley, etc). 90% of these companies will fail. Work for a pharma unless you are the CEO of one of these companies.

5. Pharmaceutical companies doing this have their informatics labs in horrible places. YMMV

Conclusion: Bioinformatics is boring. There is a lot of hype. However, if you're into it, you can probably have a good career with a pharma,  so long as you are interested in moving to New Jersey.

Needless Philosophy and unsolicited advice: Matthew Heuser above says "be willing to move ANYWHERE in the USA."  If you are a new grad, I highly recommend AGAINST thinking this way. I grew up in a small town, and I thought that I needed to go where the money was, to further my career. So in 4 years, I lived in NYC, Boston, San Francisco, and Silicon Valley proper, and worked a number of high-paying, high-stress programming and sysadmin jobs.

This was a big mistake. As soon as the job market got tight I realized that I had no support network or community to call my own if things got really bad. I had friends scattered all over the place but no one to hang out with in the here and now. All for what? so that i could say i was making 6 figures a year? I also was so focused on my "career path" that I didn't really even know how to write the type of software I wanted to write. So I took a year off to return back to school to learn a few new things, and take some time to teach myself some new programming techniques. I now have 3 telecommuting contract jobs doing exactly what I want to be doing (signal processing software, mostly audio and some video).  I don't make as much money as I used to, but I make more than enough to live close to my friends and family and I set my own hours, and I'm doing work that I find a lot cooler. So my advice is to forget the hype, figure out what you want to do, and where you want to do it. Move there, and just do your thing.

El Grumpo
Tuesday, February 19, 2002

i agree with a lot of the previous post-especially the part that most of the stuff being boring...99% of science work is pretty boring.  but-it's that 1% that's the payoff.  at least in my past experience.

razib khan
Tuesday, February 19, 2002

Having worked in both bioinformatics and web development (and web development for bioinformatics), I can't honestly say that one is more boring than the other. Both involve long stretches of drudgery puncuated by wonderful moments of synergy.

That being said, if you're not interested in genetics, it's unlikely you'll like this field -- a lot of the thrill comes from working on cutting edge and relevent stuff.

The previous poster's points about it being a field with a lot of potential for failure is true. On the other hand, what industry is actually stable these days?

Matt Christensen
Wednesday, February 20, 2002

Newgrad:

I'd lean more towards Matt Christensen's look at the situation than El Grumpo's.  While El Grumpo does make a few respectable points in his post (choosing personal happiness over who pays the most, for example), I think his view of bioinformatics is way too narrow for it to be relevant.  Anyone who thinks bioinformatics is all Perl, will exile you to New Jersey, or is all about SQL obviously knows very little about bioinformatics.  Not knocking you Grumpo, just tellin' it like it is. :)  So yeah, HE finds it boring (or rather, he finds his limited view of it boring), but I think that sticking thumb tacks into my small toe is more exciting than signals processing, so weigh your personal preferences more than other people's opinions.  ;)

If you genuinely like genetics, then consider going for it.  It can be a steep inital learning curve if you're coming from a non-biology background...

Crimson
Thursday, February 21, 2002

Of course, the moderate response is more genuine and accurate. But, this is the internet, and hyperbole is more fun. ;-)  And yes, most people probably wouldn't find audio DSP very interesting.

However,  newgrad is a newgrad, and was asking if bioinformatics looks "interesting for the long haul." Since Crimson doesn't think I know what I'm talking about, I have to jump in and say that I have more experience than I let on. My first job out of school was as a research programmer with a lab at a large medical school in manhattan for a year. I was also taking night courses in molecular biology because I was considering getting serious about the field.

When I recently went back to school (for some physics and EE courses), I worked as a part time contractor with the pharma research lab run by my relative mentioned above.

Also, relative mentioned above worked directly for James Watson at the NIH for 3 years in the early 90s. Watson is known for bizzarre public appearrances and for making statements which offend entire professions, genders, and ethnicities.  One of his favorite quotes was "The problems of BioInformatics are boring, and the people who work on them are boring. "

Anyway, what I'm getting at...is that I do sort of know what I'm talking about.  And, my opinion is that the field is better off in the hands of people with a biology background who have received some additional training in software. If you don't already have an interest in genetics (or uh...if you are one of the founders of the field) the work is not going to seem interesting in any way, other than the paycheck.

Of course, YMMV, and if newgrad thinks it sounds interesting, it is worth checking out.

Short term: there will be lots of jobs available, not necessarily just in new jersey...perhaps in fabulous locales like Palo Alto, Minneapolis, and Somerville, MA.

Long term: you WILL be working for a pharma, and you WILL be working in a suburb of Philadelphia. ;-)

El Grumpo
Thursday, February 21, 2002

Grumpo, if your recounts about how you worked for a medical school as  "research programmer" (as if any medical school is doing informatics research) in the early 90's (back when putting data into an Excel spreadsheet was considered bioinformatics) or how your *relative* worked "directly" for James Watson (who is only slightly more relevant to bioinformatics than Osama bin Laden) were to convince me of your knowledge of that area, then you'll have to paint me as unconvinced.  At best, what you do "know" of informatics is so outdated it's irrelevant anyway.  ;)

I won't even go into the insanity of such statements like how "hot bioinformatics guys" get paid less than Oracle DBAs because I don't even know where to begin with stuff like that.    I can go on and on and on about inaccuracies in your post, but I need to get some work done eventually, so I'll wrap this up. ;)

At the risk of sounding like a bioinformatics recruiter, I'll let you in on some of the interesting problems of informatics. There are pretty exciting new areas of bioinformatics in areas such as artificial intelligence, data visualization,  image analysis, and algorithm development, along with all that fun perl stuff.  But one thing Grumpo hit squarely hit on the head is the fact that you'll need to enjoy genetics or at least not be afraid to learn some fundamental things in molecular biology to do truly hardcore informatics.

Crimson
Thursday, February 21, 2002

>Grumpo, if your recounts about how you worked for a >medical school as "research programmer" (as if any medical >school is doing informatics research) in the early 90's (back >when putting data into an Excel spreadsheet was >onsidered bioinformatics) or how your *relative* worked >"directly" for James Watson (who is only slightly more >relevant to bioinformatics than Osama bin Laden) were to >convince me of your knowledge of that area, then you'll >have to paint me as unconvinced. At best, what you do >"know" of informatics is so outdated it's irrelevant anyway. >;)

Actually my bioinformatics contract job was from mid 2000 to mid 2001. For a pharma near philadelphia. You guess which one. ;-)

My initial bioinformatics job at the university was one of the first projects which used a standard relational database to store and retrieve the data, and we built a cool "application server" and custom client that ran on sun workstations so that scientists could collaborate on their work. Granted, I was just the technical dork, but I worked along side real "scientists" and would like to think that I absorbed some of what was going on.
Considering your experience in the field, it is a bit odd that you haven't heard of the project. ;-)

I could go into more details about my mystery relative, but I'm pretty sure he reads this board and I don't know if I want him to know who I am. (but he's probably guessed by now) ;-) In any case, I guess James Watson might be irrelevant to current bioinformatics research, but for a time he and my secret mystery relative approved most of the grants the government gave to molecular biologists who were doing bioinformatics style research.


I> won't even go into the insanity of such statements like >how "hot bioinformatics guys" get paid less than Oracle >DBAs because I don't even know where to begin with stuff

The starting salary for a programmer just out of school with zero experience , at a bioinformatics firm is about $40-$65 a year depending on the location of the firm and how much venture capital the firm has to burn. 

Indeed, the high level architects and scientists make more than that, but the salaries are like $80K-$120K a year with considerable experience, and most often an advanced degree in some form of molecular biology.

In contrast, someone with a year of experience and an Oracle DBA certificate can get a job with a bioinformatics start up with a low _starting_ salary of $65K and more often a starting salary of $80K. For a pharma the salaries are similar and if someone is maintaining a huge installation the salaries can be 6 figures or more.

My uncle with a phD in microbiology in contrast makes about $160K a year after 20 years of experience, 12 of them directly related to software engineering and management of bio software products. The DBA route is definitely a more efficient way to make money.

You can get this data by going to various job boards, some as generic as flipdog and craigslist. Or you can interview with firms (as I did) to find out how much they pay. Or you can ask your secret mystery relative who has been in the biz for 20 years what to expect.

It is true that I'm not a trained bioinformatician, but as recently as 12 months ago I was interviewing with bioinformatics firms and actually working in a lab, so I'm not totally clueless.

In any case, i guess newgrad will have to find out for him (or her) self what to do. I'm guessing bioinformatics is probably more interesting to some than I found it, but I'm also guessing that anyone fresh out of school is _not_ going to find the job they can land very interesting. However, this is true for most industries. You have to work a while before you have enough experience/clout to do interesting projects. I'm just saying that if you don't already like genetics, i wouldn't recommend getting into the field...because it really is more of a "lifestyle" than just a "job." But I think this is what crimson was saying too. ;-)

El Grumpo
Thursday, February 21, 2002

also, just regarding the DBA comment I made.

The big secret of the dot-com era, was that aside from some naive H1-B visa holders, the Sysadmins and DBAs on average were raking in at least 20% more than the developers of equal ranking.

I think this is going to hold true with bioinformatics as well, or any industry that is completely reliant on a certain piece of infrastructure.

Getting totally off topic, the MIT economist (and NYT columnist) Paul Krugman explains this phenomena in one of his essays... we're entering an era where skilled blue collar (and DBAs/Sysadmins are the blue collars of the IT industry) labor is going to become more expensive than skilled white collar labor. Another secret is that your BMW mechanic is making a lot more money than you are. ;-)

As a programmer, I'm not saying I like this, but it is something to keep in mind.

El Grumpo
Thursday, February 21, 2002

i've heard that about sysadmins-but i think it's because if the system is down for a whole day it's their balls that are fried-no?  no pain, no gain.

if the software is broken, debug it or patch.  unless you've got a 1 million unit distro not as catastrophic.  (that being said-i remember back in college when our servers were down for a week-maybe they had work-study sysadmins :)

on the thing about bio to computers being easier that computers to bio, that's what bioninformatics guru lincoln stein says.  he works at cold harbor where said mr. watson is now.  and watson is as weird as they say from what a friend of mine that worked at cold harbor last year tells me (now that we're all name dropping :).

oh, and about the off-topic things.  i think krugman said that stuff back in the mid-early 90s, i think in "peddling prosperity."  i don't know about it anymore.  as long as you have immigration-you're going to have people filling up blue-collar spots and depressing the wages their (chicken processing plants in iowa for instance where natives won't work in the toxic environment without remuneration that the agribusinesses won't pay).

back to sysadmin vs. developor in the bioinformatics context-my ex-roomate was the admin for the gene db at our college for fish-and he was always stressing out when it crashed or slowed or some idiot developer wrote a script that hogged up all the space.  he seemed more relaxed as a developer among developers in the more abstract work of if (geek) { return virgin} rather than hussle of % & From: Angry Technophobic Higher Up

razib khan
Friday, February 22, 2002

I have and undergad in bio and I work in IT . I am currently full time in a CS masters program.  At myschool I have been looking at, and taking electives in the Bioninformatics field.
I am not that familiar with  the business end of bioniformatics but I do have some knowledge of the things that people are doing.  For me I think the subject is anything but boring.  If you find biology or pharmacology or any related topics boring besides computer science I don't suggest the field(And vice versa).  Part of the beauty of computer science is that you can work in an alternate field that you enjoy, whether it be finance, arhcitecture, design... Personally I get excited working with something grander than just making the CEO get more profits for the quarter.  In bioniformatics you are working , in many cases , in helping solve dome disease of one kind or another.  Whether I receive 10k more in this field or that field is irrelevant to me- I guess that is the biologist in me. I also think the computer science has many cool sides to it.  Bioniformatics uses , yes databse, and databse and more database- which is great if you are into that.  Of course everyhting is opitmized to the "T" because the information is so vast and that it is always challenging. Recently I have been into the use of Neural Networks , in bioniformatics. This little niche has many powerful uses-prediction of promoter regions, secondary protein structure prediciton immunology,  and pretty much any predictable data whith unknown governing rules(many). People hae been coming up with many uses. So this is high level programming science idea (artificial intelligence) being used in conjuction with biology. Hidden Markov Models are used extensibly aswell in bioinformatics for predictin methods.  Imaging and graphics3d are used in bioinformatics and medical informatics aswell.  So I think there are many cases where one can find there computer science passion in the bioinformatics field.  Well that stuff exists all nice in academia. How does this stuff work in the real world? Are they infact used by workers? Maybe someone can help me out from the field.  The problem is once you head down one of these interesting avenues you are narrowing your field of study down to do only that one task.  Which has its benefits and definitely its drawbacks.  There are a few people out there in bioiformactis that probably can call themselves a "molecular biologlical image architect"#@ or a "Molecular Neural Network Secondary Protein Specialist", but rather a more likely job is an Oracle DBA or software enginner with a tad of bio nowledge(and vice versa fro the bio point of view).  I think that this may a more likely scenario where you here the words " you're hired." thoughts?

Steven Sanborn
Tuesday, January 14, 2003

I personally am an MS in CS, a recent grad and am working in Bioinformatics. Following are some of the things I have realized in past one year or so:

1) As a Bioinformatics developer, you HAVE to delve into genomics, proteomics and somewhat into molecular biology too, if you want to know what you are doing.

Having said that, it is not too hard. It is easier than learning Java and C++. Just a little effort and on-the-work  guidance from the right people is enough.

2) Most of the "bioinformatics programmer" positions are in academia.
Academia, for IT people not involved in research, can quickly become quite boring. Salaries are lower (15-33%) than industry, no bonuses and increments, there is not enough room for growth and for most part, you are the only IT person on the projects you are assigned.

Academic job, IMHO, is best for a laid-back lifestyle if you are not looking into doing research.

3) Bioinformatics skills, for the most part, are transferable and hence it is easier to switch jobs if you have to. But don't expect to make as much money as your friends in industry are making. It is more or less a flat curve, tied with grants coming in and not to inflation/economy/performance.

4) Future? Well, it is science and you never know. But for about a decade more, it should keep on kicking fine.

My four cents.

Soham Mehta
Monday, February 23, 2004

*  Recent Topics

*  Fog Creek Home