Slashdot Mirror
Computational Simulations of E.coli
Gearoid_Murphy writes, "BBC news has the story of a scientist who has been using computational models of bacteria to advance our understanding of actual bacteria — a step towards simulating fully fledged organisms in virtual environments and potentially an extraordinarily powerful tool for medical science."
BBC news has the story of a scientist who has been using computational models of bacteria to advance our understanding of actual bacteria
So the story is a simulation that actually simulates what it's simulating? (Isn't that what a simulation's suppose to do? Model and echo what happens in the real world)
How stimulating.
These posts express my own personal views, not those of my employer
We'll start with bacteria and move our way up to humans! ;)
Hopefully this would eventually allow risky medial treatments to be simulated before they have to be performed with a scan of the patients physiology as a reference.
Shh.
... a fully 3D representation of the work-counter of a kebab shop, where the E-Coli simulation starts?
Hence, discrepancies between what the scientists see in biological experiments and what they see in the simulations allows them to test the models. If there is a mismatch it suggests the model is incorrect and needs to be refined.
Wow, that kinda sounds like.. umm, what's the word I'm after here, umm, science, yes, that's it.
How we know is more important than what we know.
That article pretty much sucks for communicating anything that the scientist is doing. I won't remedy that, but I'll say this: I audited a biophysics class two semesters ago, and the astounding complications of what goes on in cells was a real eye-opener. Of course, I'd learned about chemical pathways, and mitochondria, etc., before, but the class showed how damned complicated and *fast* everything is at the cellular scale.
The professor used this analogy: think of filling a football/soccer (your choice) stadium with ping-pong balls, and paint just two of those balls orange. Then hire some bulldozers to push the balls around randomly and continuously for several decades. How often will the orange balls collide with each other? Once a week? Once a month? Once a year? Maybe only once in a decade? Now envision the stadium scaled down to the size of a cell, with the ping-pong balls now being your average-sized molecule important for some process (chunks of amino acids, say). These will be moving around randomly due to Brownian motion, chemical gradients, etc. How often will two given molecules interact? Probably several times per second. THAT's how amazingly extreme cellular processes are.
It's that sort of analogy (sorry it wasn't about cars, but we could probably work those in somehow) that the article should have had. This stuff is complicated, and requires VERY efficient computation. Kudos to the researchers, and pfft! to the author of the article.
Wow, figures. I actually used TFA for my Cambridge application -- nice to see it appearing on /. albeit slightly later than I expected. Really though, Cambridges Computer Science Lab (funded by our friend Bill Gates, among others) is doing some amazing things. Check out their website at http://www.cl.cam.ac.uk/.
games journalism blog
Frankly, I like that image better than most of the ones propogated by common religions.
I wouldn't mind Heaven if I could go grab a beer and some bowling with God.
Fill in your four or five-letter word of wisdom here _ _ _ _ _.
You don't need to know *WHY* Guinness tastes better than Budweiser, all you need to know is that Guinness *DOES* taste better than Budweiser. Thus you can spend your money on quality beer right from the outset.
When our name is on the back of your car, we're behind you all the way!
I've had E.Coli before, but I'm not quite sure what a computational simulation of severe diarrhea would look like; maybe Microsoft code?
Gravity is a contributing factor in nearly 73 percent of all accidents involving falling objects. -Dave Barry
Very interesting comment. For once, an insightful comment from an AC.
I dont know what they modelled but I know why they modelled it. So that the California Spinach farmers can claim, "No E Coli was actually harmed in filming this commercial".
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
They had articles like this ten years ago, that some grad student somewhere simulated an "actual bacterium" on a PC. The topic is useful primarily as a graduate student's thesis topic, along with "robotic simulations".
a step towards simulating fully fledged organisms in virtual environments..
Yeah but it's like, a really small step isn't it? You couldn't do anything useful with it except maybe simulate Dubya's cerebral activity, and that's not very useful at all.
And besides, didn't you hear: Virtual Environments/Machines are going to be banned on Vista!
As soon as some beverage company decides to model human taste buds in order to determine how they can modify their beverage to make it more appealing to consumers without actually modifying the process by which they brew their beverage, brewing a new batch, and conducting taste trials. Simulations are likely to cost less than actually making and testing the modifications to their recipes or processes.
I don't know this research, and the article doesn't really say anything at all, but as a grad student who has done a lot of cell modeling research, I like his approach of limiting the model to something very simple and easy to verify. We are a long, long way off from "simulating fully fledged organisms in virtual environments". Probably not in our lifetimes. You just have no idea how complex even E. coli is until you study it, and if you have, you'll understand how primitive and limited our models are.
If the simulation framework would already exist, you would be right, but in the particular case you mention you're probably not: you'd have to first find out which receptors in the tastebuds are responsible and what their structure is. To be honest, computer simulation is still far away from such amazingly complex tasks.
Furthermore changing a beverage can be done fairly cheaply, you just need a small test setup to do the brewing, and then do batches after each other, you could try a lot mixes in a month. Then for tasting you could use an "artificial nose", or just do it yourself.
Anything that contains a molecular mixture with loads of different chemical interactions (I'm not even considering reactions here) is still an immense task for simulation, as all the parameters you put in to it have to be tested individually. Current simulation has strengths, e.g. in material design to test high-pressure extreme force conditions, that would be either impossible or very expensive to perform. In the case of beer-brewing probably the most in reach of simulation would be to optimize foam stability, although even that would be a pretty difficult task!
molmod.com - computing tips from a molecular modeling
TFA (that is customary not to read) may be short on substance but it can also be said "I can pull results out of my arse", science = observations + models, you cannot judge any result without access to the observations and a basic understanding of the model.
And did you exchange a walk on part in the war for a lead role in a cage? - Pink Floyd.
Reminds me of this book (I think I got the right name, read it a couple of years ago.)
Hard-scifi, check it out!