Cold Spring Harbor Laboratory bioinformaticist Lincoln Stein is a cofounder of the annual Genome Informatics conference, held Nov. 1-5 at Cold Spring's facilities in Long Island, NY. While no longer involved in organizing the event, Stein spoke there this week to provide an update on one of his pet projects, GBrowse.
The latest version of the open-source browser, originally developed in 2002 for the Generic Model Organism Database project, is part of the continuing aim at the Stein lab to produce reuseable software components for model organism databases. Recent enhancements include draggable tracks, pop-up balloons, and AJAX technology.
Stein, who is moving to the Ontario Institute for Cancer Research full-time in 2009 (BioInform, Aug. 24, 2007), said that the upcoming version 2.0 of the browser was inspired by Google Maps and had to undergo a comprehensive restructuring by himself and the GBrowse development team: Ian Holmes from the University of California at Berkeley’s Department of Bioengineering; Mitch Skinner from Simon Fraser University in Canada; and Scott Cain of CSHL. GBrowse 2.0 is slated for release in the first quarter of 2008.
Stein’s other projects include Reactome, a database of biological pathways; the Arabidopsis Insertion Tagging Database; Wormbase, a model organism database for C. elegans; and Gramene, a genomic database for rice and other grasses. He is also a core BioPerl developer and has authored several books on Perl.
BioInform spoke to Stein briefly between talks on the last day of the conference about GBrowse and why he’s switched from studying mice to C. elegans in the lab.
You've spoken to us about GBrowse before, so what makes you feel that all these bells and whistles right now are giving it an extra kick?
A lot of bioinformatics activity is decentralized; [GBrowse is] a great visualization tool that is accessible to small labs as well as large labs that start out small and scale it. It's compatible with all the model organism database identification systems and data model and management tools, so it's a good choice for someone creating a genome-level informatics resource at a lower [point] of entry than other big browsers. I think the community supported it; and the community's response to … these new features, I think, is very good.
Right now, you are going [to the Ontario Institute for Cancer Research] one day a week; is that right?
And you are [at CSHL] four days a week, and then in the beginning of the year you will be there full-time?
A week a month beginning in January; and then I could [move] to three days a week in the spring; and then full-time in January 2009. As soon as my wife finishes her thesis work we will relocate.
Mind if I ask what she does?
She's a neuroscience PhD student here at the Watson school for Biomedical Sciences. She works on the “behaving rat” — rat behavior.
Any thoughts when we will be able to do away with rats and mice in the laboratory?
And replace them with computer models?
Well, I hope it happens sometime, but it's not going to happen soon because so little is known about physiology; about how things like the nervous system work; the immune system. There’s just so much that's unknown. It’s going to be impossible to move away from mammalian [model organism] systems.
I did work with mice for my graduate work; I found it very emotionally taxing to kill the animals and give them pain, so I stopped. Now the lab work I do is C. elegans, where I don't feel much moral compunction. But you know, it's a personal choice for every scientist whether they work with animals.
What do you think of Perl versus Ruby these days?
Ruby is really looking terrific. If I were to do it over again, I would consider Ruby very strongly. It's my development language of choice. It doesn't provide enough advantages for me to go through the pain and effort of retraining [myself]. …
On another note, I noticed that everybody who had a presentation [here at the Genome Informatics conference] seemed to be hiring.
Is the field booming? And if so, why do you think that is?
It's certainly booming again. It boomed a lot in the late 90s. There was a bit of [connection to] the [dot-com] tech boom [then].
[A]ctually, [now] more of what you are seeing [is] more of a statement about what's happening in the high-tech industry outside biology than what's happening in biology.
During the tech boom, all the talented software developers and mathematicians went into the Internet, into Internet startups and financial companies. Then there was the balloon popping, so all of that talented group of people went into academia. There was a glut for a while, and now there is a second Internet boom again and it's grabbing the people away again.
At this summer’s ISMB, Temple Smith told me that a lot of people at the informatics level might not necessarily understand the biology for which they are creating tools. And he saw that as a bit of a problem.
It's always a problem if bioinformatics and computational biology are 'ghettoized.' They need to be integrated; the most successful sites for bioinformatics have a strong integration between bioinformatics and the lab. And so, I have seen an effect in which institutes that were founded for just bioinformatics [and] didn't have a wet lab component went off and did things that were very pretty, but not what the biologist, the experimental biologist, actually needed.
So that's what he's saying; I agree with it. But the successful organizations have both groups of people sitting together in the same building.