?

Log in

No account? Create an account
 

Talking about Biocomputing - F*cking with Clusters

About Talking about Biocomputing

Previous Entry Talking about Biocomputing Oct. 3rd, 2014 @ 11:01 pm Next Entry
Dear LazyJ,

I'm giving a 20 minute talk about computers and biology to a potentially large, very mixed audience in a week(!). I have not yet written said talk. I'm trying to narrow down what parts of the intersection of bio and computing to cover, and trying to get the point across that even bioinformatics has problems that involve science for the CS side (as opposed to just rote system building / data analysis). I also want to give a broad overview of the biology and computing research intersection.

What sorts of topic would you expect to see in a talk titled "Computers and Biology: Beyond Gene Sequencing" that has been crammed into a Data Science track (for absolutely no good reason. It's like they didn't even read the abstract :/).

In other news, it's jobhunt season. I had a great talk today with a guy I knew a bit at Mudd, but never really spent time talking to. He's a CS prof in a department that's hiring. I know nothing will come of it, but it's still a tiny bright spot in a long bleak tunnel.
take a penny
[User Picture Icon]
From:derakon
Date:October 6th, 2014 04:20 pm (UTC)
(Link)
Sounds like you're driving the CS side of things a bit? That is, you're saying "Here's how CS and Biology intersect to make interesting and difficult problems on the CS side" or something like that.

Protein folding is a classic hard biology-CS problem -- you have a collection of proteins that are allowed to rotate in certain prescribed fashions, and you try to find the configuration that minimizes the total energy of the system. You could talk about various optimizations to the automatic folding programs, Folding@Home, as well as the fact that it's been made into a videogame (that has made a few scientific discoveries).

There's a lot of iterative optimization problems in microscopy, where you have a dataset (a collection of images), and you're trying to find a function that best fits that data. I guess that's "just data analysis", but stuff like image deconvolution and compressed sensing can be very cutting-edge.

I'll see if I can think of any more good examples. Technically I'm in the CS/Bio realm but mostly I do application development these days, so I'm a bit more disconnected than I used to be.

Good luck with the jobhunt!
[User Picture Icon]
From:avani
Date:October 6th, 2014 04:33 pm (UTC)
(Link)
Thanks for reminding me about protein folding. That's a great example.

What I think would be most awesome is more examples of taking modalities in CS and applying them to bio research. For example, if we can show that the brain does the same sort of image deconvolution that vision systems do (which is something my lab kind of works on, but it's a stretch)
[User Picture Icon]
From:derakon
Date:October 6th, 2014 05:19 pm (UTC)
(Link)
Hmm, unfortunately I don't have as much breadth of knowledge of CS research as I used to. I'm pretty sure there's been research showing that ants execute graph search algorithms, or something along those lines. But that's more of a "hey, that's neat" kind of discovery IMO, also most people don't think of entomology when they think of biology.

...which is dumb, in retrospect. Would stuff like analyzing how dragonfly wings work in order to build better flying robots count? Because people totally do that. There've also been studies of how cockroaches walk; in general studying insects is great for making smaller, "smarter" (more efficient/effective) robots.

(Also, re: protein folding, it's not the proteins that rotate, it's the components in the protein, my bad.)
[User Picture Icon]
From:sithjawa
Date:October 6th, 2014 09:02 pm (UTC)
(Link)
Any chance the talk is gonna be streamed? :D

What topic would I expect to see covered? There's really quite a lot of different ways of approaching bio/computing. It depends on the audience and aims. Since you say it's a broad/mixed audience and your title is "computers *and* biology" not "computers *in* biology" or "biology in computing" I'd expect that you might touch on the ways that biological systems can inspire new approaches in computing (genetic algorithms, neural nets would be the classic examples, but I feel certain there are newer examples I can't think of at the moment, especially in robotics). derakon's most recent comment seems to be on similar lines. Dunno if that helps at all/is even in the right direction.

Now I'm reminded of that one TED talk about the slime molds. Definitely more on the "out there" side, but it's interesting.

http://www.ted.com/talks/heather_barnett_what_humans_can_learn_from_semi_intelligent_slime_1?language=en

While I was looking that up I found these people building a slime mold powered chip. Science is weird...

http://phys.org/news/2014-01-slime-molds.html
http://www.phychip.eu/
[User Picture Icon]
From:rose_garden
Date:October 7th, 2014 03:50 am (UTC)
(Link)
Today's xkcd is related
(take a penny)
Top of Page Powered by LiveJournal.com