4 ms·

We need more computer scientists in biology and biomedical research

6 points by Protostome 8y ago

bufferoverflow 8y ago
Who is this "we"? When I graduated with a degree in bioinformatics, the offers I got were ridiculous, the best one was $54K near NYC. And I was already a published scientist with years of programming experience at the time. At the same time salaries for web dev were great, so I stayed in web dev.
- Protostome 8y ago
  "We" - the scientific community. Coming from the academia, the need is certainly stronger there though. I graduated ~7 years ago and was also a published author. at the time. The salaries I was offered for bioinformatics positions were lower than salaries from companies from sectors like ad-tech or security, but not by a lot. For me , salary is not the most important factor in the decision and I rather work on interesting and important problems than earning 20% more. (although 54k is ridiculous for a published scientist).
- iso1337 8y ago
  Yeah, that and the incentives are not there for programmers. They are often treated as the support, not given proper training and managed by people who has no idea how to develop software. This is slowly changing but is still bad.
maxander 8y ago
> Computer scientists should be involved in the R&D process from its very beginning, not just as sophisticated data analysts. Starting from the design of the wet-lab experiments — where the data is being generated. This is because we, as computer scientists understand that data is the most important resource. Your model, whether it is an incredibly sophisticated CGAN or a logistic regression is as good as your data. This is the key point in the article. There's lots of bioinformatics and computational biology work being done, but the vast majority of it winds up being of questionable use- either because the researcher didn't understand the biological system well enough to make an appropriate model, or because the biologists at the bench didn't understand the analytical problems well enough to collect the right kind of data. The typical model of research is "design experiment -> collect data -> analyze data -> results", and when each step is being done by different people, with different training and no input into each other's work, it's tremendously wasteful. It's the science equivalent of the waterfall method, except that instead of technical debt you wind up with dubious publications.
arandr0x 8y ago
This was the career I wanted, but I'm not willing to work in academia. I actually have a biochem degree and 10 years of work experience as a software engineer. The point should be rephrased : "We need cheap programmers in biology and biomedical research in San Diego and Boston to blindly obey senior cancer researchers with fat pockets, and we can't afford the market rate for a good programmer or data scientist in those two arguably very pricey locales, and we don't want to hire remote workers, even though the job is desk-based. Ideally they should be willing to be graduate students, that's easier on the budget."

6 points by Protostome 8y ago

Who is this "we"? When I graduated with a degree in bioinformatics, the offers I got were ridiculous, the best one was $54K near NYC. And I was already a published scientist with years of programming experience at the time. At the same time salaries for web dev were great, so I stayed in web dev.

Protostome 8y ago
"We" - the scientific community. Coming from the academia, the need is certainly stronger there though. I graduated ~7 years ago and was also a published author. at the time. The salaries I was offered for bioinformatics positions were lower than salaries from companies from sectors like ad-tech or security, but not by a lot. For me , salary is not the most important factor in the decision and I rather work on interesting and important problems than earning 20% more. (although 54k is ridiculous for a published scientist).
iso1337 8y ago
Yeah, that and the incentives are not there for programmers. They are often treated as the support, not given proper training and managed by people who has no idea how to develop software. This is slowly changing but is still bad.

maxander 8y ago

> Computer scientists should be involved in the R&D process from its very beginning, not just as sophisticated data analysts. Starting from the design of the wet-lab experiments — where the data is being generated. This is because we, as computer scientists understand that data is the most important resource. Your model, whether it is an incredibly sophisticated CGAN or a logistic regression is as good as your data. This is the key point in the article. There's lots of bioinformatics and computational biology work being done, but the vast majority of it winds up being of questionable use- either because the researcher didn't understand the biological system well enough to make an appropriate model, or because the biologists at the bench didn't understand the analytical problems well enough to collect the right kind of data. The typical model of research is "design experiment -> collect data -> analyze data -> results", and when each step is being done by different people, with different training and no input into each other's work, it's tremendously wasteful. It's the science equivalent of the waterfall method, except that instead of technical debt you wind up with dubious publications.

arandr0x 8y ago

This was the career I wanted, but I'm not willing to work in academia. I actually have a biochem degree and 10 years of work experience as a software engineer. The point should be rephrased : "We need cheap programmers in biology and biomedical research in San Diego and Boston to blindly obey senior cancer researchers with fat pockets, and we can't afford the market rate for a good programmer or data scientist in those two arguably very pricey locales, and we don't want to hire remote workers, even though the job is desk-based. Ideally they should be willing to be graduate students, that's easier on the budget."