Slashdot Mirror
Supportive Courses for Bioinformatics?
Per Christian Henden asks: "I`m aiming for a masters degree in bioinformatics, and I`m uncertain which courses would be good to follow (and my counselor doesn't know, either). There are of course some courses that 'belong' to this degree, and I`ll take those, but I get to choose a number of additional courses. I want to ask people working in bioinformatics 'What (CS) subjects are important or especially useful in bioinformatics?' I`m planning on choosing 'Large Datasets', 'Parallel Programming', 'Image Classification', and 'Subsymbolic AI', because I think those are important, but I`m really not sure what is useful or not in real life bioinformatics." Other Ask Slashdot articles, which have touched on bioinformatics, have dealt with magazines, books and graduate schools.
I work for a Bioinformatics company. Biology Biology Biology Biology!!! We have, and they are not doing as well, several people who with more of a CS background. They don't understand what it is that they are writing programs/algorithms about. We have found it easier to take Biologists who know computers and get good work from them than to do the converse. Make sure you get a VERY good grounding in basic biology. If you don't know what the data you are looking at mean, in the biological sense, then you will make the same mistake a lot do. Just because an algorithm is cool doesn't mean it makes sense. Only by understanding the Biology can you understand the difference.
Gavin Fischer
Your first problem is asking a counselor.
Find a professor at your University doing this, or interested in this, and get his advice. Better yet, start working for him.
That's just generally good advice for anyone who wants to go to grad school in science or engineering.
subject line says it all.
No, seriously, this is a good book.
668: Neighbour of the Beast
First, it would help if you told us what they're going to make you take for the program.
Second, what is your background. Do you have a CS background? Biology? English?
As for the courses of those Large Datasets is probably the most useful. Image Classification? Depends. If by 'bioinformatics' your actually talking about Biomedical IS (things like Query By Doodle for retrieving X-Ray information and so on) then yeah, but working on sequence data there aren't any images to classify. Maybe for some of the labeling experiments, buts thats more a problem in segmentation not classification. I'd call it less useful.
I don't even know what you mean by Subsymbolic AI (search? more image analysis?). Depending on what you meant by that it could be useful. There's a good deal of crossover between text analysis and sequence analysis.
Parallel Programming? Not so much I'm thinking. Most algorithms you're interested in are trivially parallel so at most you need scatter-gather type parallel programs, unless by Bioinformatics you mean Computational Chemistry then you could run into trouble.
If you're background is Biology you'd be best served brushing up the Math skills (Statistics and Applied Mathematics, Mathematical Modeling or Biomedical Modeling if you can find it). Computer Science, Math (depending on your program) and Biology. Math? Biology.
English? Math, Biology and some intense programming experience.
You can learn enough about programming language later on to write programs to do what's needed. Learn about as much biology as you can, as well as as much as you can about algorithms (e.g., phylogenetic methods, alignment methods, micro-array analysis methods, and lots and lots of statistics). In your spare time, learn Perl.
social sciences can never use experience to verify their statemen
I am constantly confused by what is bioinformatics. I always thought it was data relating to organisms. That's a pretty damn broad definition of it, however in my experience, using bioinformatics in a search engine rarely gets what I want.
Population pharmacokinetic and pharmacodynamic modelling is a huge field. I have a masters degree in it and as a pharmacist I can tell not a lot of knowlegde in biology is really required (if that is a problem for you). If you want cutting edge stuff there are clinical trail simulations my mentor is working on, which is I guess AI(?)
Needless to say, anything pharmaceutical commands huge amounts of money, so if this is what yo uare looking for, check out names like Lewis Shiner, Stuart Beal, Steve Shafer and Nicholas Holford.
"I just can't sit while people are saying nonsense in a meeting without saying it's nonsense" J Watson, Sci Am 288:(4)51
It depends what you mean by bioinformatics, which in itself is a broad and fairly uninformed decision. Most of the time these days people mean molecular biology, so for a good overview of what it means in molecular biology have a look at this. Then it really depends on what you want to do. Do you want to administer the systems, identify new software and try to maintain and improve a toolset? In that case you need some computer science. Also it's also a hell of a lot easier to be a self taught computer person than it is a biology person. If you want to do mainly lab work, then you really need the biology, but doing all of your spare classes on computers on basic algorithms, basic programming, a management / design subject and would be good for getting into bioinformatics. Perl has quite a strong presence in molecilar biology bioinformatics and is a handy "glue" language, as a lots of data passing through different specialised processes, and helps to teach you stuff about libraries, how to use the world wide web and such like. I'm personally more interested in the application of computers to ecological questions, and if that's your thing, then you need GIS and a solid statistical background, and not to be too scared of maths. You also need to have a handle on basic biological processes like sex and death, and evolution is very important. This becomes more of a task of managing extremely rich data sets. If you have a computer background already you need to study biology, and get some hands on expreience on a good research project that will stretch your computer skills. From a biology background it's slightly different. Try to learn how networks work, how to read source code and how to glue programs together. That's my exprience so far anyway.
"...we should just trust our president in every decision that he makes and we should just support that." B.Spears 2003
Univ Calif, Irvine's CS dept has a Bioinformatics specialty/emphasis for graduate students.
UCI's Institute of Genomics and Bioinformatics
UCI ICS Informatics in Biology and Medicine
They just got a $4.2M grant from the NIH to strengthen and consolidate their Bioinformatics group. Maybe they'll offer scholarships.
I work for a huge Bioinformatics department at a major pharmaceutical company. One thing you need to understand is that the lion's share of Bioinformatics would probably be considered "run of the mill" stuff as far as the rest of computer science is concerned. The biggest problem in Bioinformatics is that our IT people lack sufficient understanding of the underlying biology. If I had to pick an ideal candidate for a position with us, I would look for the following (in order of importance):
Biology background that includes some of the following undergraduate courses: molecular biology, genetics, evolutionary biology, biochemistry/biophysics, virology, & microbiology
Computer background that includes many of the following: Perl, Java, relational databases, XML, Visual Basic (lots of stupid stuff uses it) & general UNIX skills
Preferably some kind of lab experience
Any background in statistics (the more the better)
What we really need are people capable of integrating large amounts of data (DNA sequences, protein structure info, gene expression results, biochemical pathways, clinical data, etc...). This stuff all needs to tie together somehow. While this doesn't require massive programming heroics, it does require an attention to the underlying science or else the wrong assumptions can be made and a total disaster created.
Now, that's not to say that people with experience in areas such as artificial intelligence, image analysis, or natural language processing wouldn't be assets, but to be honest, we're just not at that point yet. Some actual Bioinformatics companies and organizations (TIGR, Celera, Affymetrix, Rosetta, etc...) may (and probably do) have a need for these specialized skills more than us. But do yourself a favor and take as much biology as possible. It's invaluable to my "office karma score" to be able to discuss the latest on Slashdot with our programmers, and two minutes later talk about an article from Nature with our biologists.
Depending on exactly which area of bioinformatics you're looking to get into, statistics may be an important aspect and is often overlooked. Not just your basic stats 101, but understanding what hidden Markov models, Monte Carlo simulations, and Bayesian analysis are good for.
:p.
I didn't take enough stats and have no real good grasp of the above
-h3
I`m aiming for a masters degree in bioinformatics, and I`m uncertain which courses would be good to follow
So take a Master's degree in Bioinformatics. What's to be uncertain about?
Learn biology, and specially cybernetic biology, the modern stuff (post-1970s) that most schools won't teach. You probably will have to go to Europe, or perhaps Latin America to find such courses, though.
``L'imagination au povoir.''
I work for a market research firm that focuses exclusively on the bioinformatics market. With a bachelors degree in Mechanical Engineering I am the worst educated employee of the company if you don't count the President/co-owner with his bachelor's in history.
The main problem we have is that all our employees, regardless of field, need a strong scientific education (there's only 15 people and the company is profitable, so there is a lot of cross-utilization). Even the marketing folks come from a lab background, so they understand what is going on. I am their resident IT geek and in the last year I have been forced to absorb more about biology tools and techniques that I would have even dreamed of, and that's without even trying.
Pedro
----
The Insomniac Coder
Nope - Not Perl. C++ and Java. Perl won't get it done. Data set sizes are too large and speed too important. Crap code won't sell either (as that feature does shine through), and we are distinctly a 'for profit' entity.
I would agree that the best stuff comes out of a tight collaboration between developers and Biologists, but both parties involved need to know the others language. We have seen many of the folks coming out the Bioinformatics programs that are just CS people who thought there was money to be made in Bioinformatics... This way lies darkness. If they don't understand the Biology, and for the most part the Bioinformatics people we have dealt with don't, then they are not as much use as someone who understands both. People who are interested in code frequently get stuck in the "because it is cool" or "Because I can" type of development, in stead of stepping back and wondering "Why would you do that" or "Does that tell me anything?"
I agree though- you need both.
Gavin Fischer
University of Illinois at Chicago actually has Bioinformatics programs and their curriculum is available on their website.
When proteins are on the line, you need expert, professional help.
There are no karma whores, only moderation johns
I have a few papers published in the BioInfo area and so alittle experience.
i) Biology or BioChem
ii) Maths or Statistics
Reaons: Bioinfo is about handling large data sets of Biological information. Therefore unless you have a good understanding of Biology and/or BioChem you haven't a clue what to actually do with the data.i.e. what are you looking for?
Once you have access to the data and want to extract significant meanings from it you'll need somewhere along the line to use maths/stats to figure out what the data is saying to you. I hate to pour cold water on the CS people but if you have a good grasp of Bio and Maths the programming is easy. 90% of BioInfo programs are written in Perl or Python because it easy to code in those languages for non-cs people. Also Perl/Pythan handle data i/o well which means whatever database you have you can nearly always code up a Perl/Python script to extract the data from it.
We've had life sciences people coding up their BioInfo scripts in 1wk to produce results. After a month or two they don't need any help at all. Their programs may not be efficent but they cerainly work. On the other hand the 2 people in our lab (out of 16) who have Stats training are overworked in helping the other 14 analyse their data to find significant trends.....
Just my 2p worth
Jon
My ZX81 could do that, but only on a good day!
I recently visited the Human Brain Project yearly conference at the NIH in Bethesda Maryland. It answers all your questions. There was one computer person giving a talk for every three researchers' talks.
The big buzzword this year was Bayseian Filtering. People were using it to model probabilities that genetic sequences would correspond to: Behavior, electcrical signals in specific cells, pathology.
Lots of people were using Java. a few more people using PHP, Python and Perl... no mention of C# or other microsoft stuff. Two projects were using Qt extensively, although they were a little disapponted with something to do with Jpeg image stuff on OS X. I got the impression that they didn't see Qt as being perfect.
They're mostly Mac people And there are a _lot_ of linux enthusiasts. I had dinner with a handfull of people who seemed like 50+ year old administrators. When the subject of Linux came up things got animated. One gentleman runs every SuSE on every desktop in his department. (using Crossover plugin for MS office)
Lots of people are putting together very large databases, and trying to model lots of complex interactions. Others were trying to standardize cross-database communication.
Henry Markram from Switzerland stole the show with a Neocortical microcircuit database. He has measured electrical signals from thousands of living brain cells, traced the cells using Neurolucida, and found corresponding factors between genetic sequences and patterns of electrical activity.
It's clear there's still a lot of excitement (and a little money) in bioinformatics. The important thing is to make sure it's being driven by science.
Almost all of the work that's being done (in this field) is being done in Academia by cheap graduate students... so the complaint is that the programmers rarely stay with the project for more than a year or two.
-Jim
Celebrate Excellence!
Not knowing how to assess significance of your results can cause you a lot of grief, either if you miss important results, or assign importance to non-signifcant results.
In the second tier, it is crucial to have a good grasp of the biology itself - and here it highly depends on what you're researching. If you deal with genetics, molecular biology is critical; the more the better.
Also, take some machine-learning courses if you can (those that deal with classifications).
The software-writing part in computational biology is less important, and as long as you feel comfortable in whatever programming language, it'll probably do. The difficult part is knowing WHAT to program, which data to use, and figuring how to extract significance out of it.
I have been working in bioinformatic for years now and can say conclusively, "WE DON"T NEED YOU!!". There are already a ton of CS jerks that don't know any bio and are just looking for a job. The very fact that you are asking this question indicates to me that you will never be able to make any meaningful contribution to the field. Move to India to be a software jabroni. Bioinformatics needs skilled professionals NOT YOU.
I hope i can shed some light into why i've enrolled in Bioinformatics...
:/) I'm still facinated by genetics and the human (et plantae, animalia) genomes. I think there is HUGE potential in genetic engineering... However, i know where the work is, and where the discovery really is.
:/)
d =1476685
I chose Bioinformatics Undergraduate course at Birmingham (http://www.bham.ac.uk) because it was a sensible blend of Biology (something I'm 'good' at msyelf and supposed to be inherantly good at with my father being a GP, mom a nurse and grandad a health inspector) and computers, more specifically, perl and databases.
I've found a love of genetics recently, since embarking on my AS levels two years ago, and today, the eve of my A2 level exams (AQA, 1:30hrs back to back
Most people say that the real discovery is at the lab level, but after years of "oh well, it works" and "oh well, it didn't quite work this time, maybe i used too much ocetic stain?" to "Damn it! I've just thrown away the supernatant and not the fsuking sediment!!! ARGH!" i've come to believe that lab experiments are really left to those with a different type of enthusiasm i've yet to acquire. While i think experiments can be rewarding at times, total discovery isn't always as fun.
The blend of bioinformatics with perl was something of a godsend when i started reading up on the subject. I've seen conferences advertised, fm and sf projects on bioinformatic processes and a oreily liturature.
Tbh I found the course at Bham by pure accident. Somehow a Biosciences leaflet came through the door, and on it was a limited place course with a grade boundary i could squeese into. I emplore you to do some of your own googling like I have.
I found stats like, "Bioinformatics is estimated to be worth $40Bln in the next 3 years". I know the biotech industry is very down at the moment (i do Business aside from Biology, and had a portfolio in some Biotechs for a module), but i can only envisage it booming in a few years as processing power increases and genetic probing is refined more and more. Only in the last 2 years we have seen the manufacture of mass probe machines, and they're being improved on year on year.
Ultimately i decided that i'd rather do something that would not only enhance my perl and database skills - which are currently limited only to small-time webapplications, and aimed at small scale systems - but reward me careerwise. Bioinformatics includes the necessary skills to migrate into CS or into other Biological Sciences, such as Genetics (tho the natural progression is Genetics to BioInf). I see Bioinformatics as the only "real" next big thing on the tech and biological market. While it's true the pharma conglomerates are conducting dubious business, it's this business that's been generating *MASSIVE* revenues per product; if I am able to manage a new product, such as an artificial enzyme or protien, from conception to the market place, i can only see big $$$s as my reward (read as pounds, i don't have unicode installed
Do read, http://www.economist.com/displaystory.cfm?story_i
My dad put me on the the fact that Bioinformatics doens't have the technologcical-bust stigma CS is gaining. I'm coupling this with a top 4 UK Uni, and looking at postgraduate conversions to almost anything if I can get a 2nd class... That said, I think there are going to be interesting implications in LAW and Bioinf soon. An LLC conversion course to specialise on genetic, information and biological law is not a totally absurd plan at all! (i have convinced myself that anyhow)
I know one thing for sure is that whether or not the Biofinformatics world kicks off when i graduate in 3/4 years time (assuming i make the grades, which is oh-um-err right now) I'll come out with experience in CS and Biology and a broadrange of skills...
Goodluck.
Matt
http://bio5495.wustl.edu/6 82/
there are some books mentioned recommended by the professor at http://www.genetics.wustl.edu/bio5491/
you will doubtless be using some pretty heavy duty statistical analysis, check out a syllabus from my undergrad school on statistical genomics.
http://www.bscb.cornell.edu/Homepages/raaz/BTRY4-
Good luck!
For the question about subsymbolic AI:
The field of AI is divided into subsymbolic and symbolic AI, where the symbolic AI guys are working with formulas in logic and programming in prolog, typically, and the subsymbolic AI guys are programming in Lisp and working on neural networks, genetic algorithms and genetic programming.
Your not going to get far in bioinformatics coming from the CS side. You need a good background in biology and/or biophysics. There is a top notch program at Johns Hopkins which is designed for CS people who want to work in computational biology. (They have a tough time attracting CS types because the pay is lousy, so they are always on the lookout for good people). Check it out at: http://www.jhu.edu/~ibr/bwf/index.html Good luck!