| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Help Conquer Cancer
Thread: The Limitations To Bioinformatics |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 8
|
|
| Author |
|
|
Thomas Klauset Aurdal
Advanced Cruncher Joined: Apr 27, 2007 Post Count: 69 Status: Offline Project Badges: 
|
What are the limitations to bioinformatics in general and the limitations to the bioinformatics used in BOINC at WCG?
Are there some biological processes that can't be simulated however much computational capacity you have access to? Are there some biological processes in systems biology that have to go through experiments in the old trail-error way? I would like to see some scientists answer the question in a comprehensive way. Maybe I have to refine the question but I think it's possible to understand. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello Thomas Klauset Aurdal,
It has been a long time since we posted anything on this topic. The first 2 posts in this thread touch on the subject: https://secure.worldcommunitygrid.org/forums/wcg/viewthread_thread,2066 In general, scientists are experimenting with cheap computational biology, but are still pouring vast amounts of money into 'in vitro' lab work, which tells us what their general opinion is. Help Conquer Cancer ( http://www.worldcommunitygrid.org/research/hcc1/overview.do ) is trying to use the computer to speed up lab work on determining the structure of crystallized proteins. Lawrence |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
My understanding as an applied mathematician is as follows.
----------------------------------------The only process that couldn't theoretically be simulated is one that we don't have a model for. Strictly speaking we don't simulate processes, we simulate a model that is reasonably accurate at describing the process. Experimentation is necessary to determine if the models are accurate, and how much so. Our understanding is accumulated in the models that we build to explain processes. Data that doesn't fit a model which we expected would fit it, or for which we have no expected model is probably the most interesting, and the process that generated it is not able to be simulated until a model is found. Of course some models which we may know well, may not be able to be simulated due to computational complexity. Models may be simplified to make them more tractable for computation. Also a lot of what is crunched is not simulation, but data analysis on the data generated by simulation or experiment. [Edit 1 times, last edit by Former Member at Feb 21, 2010 8:02:25 PM] |
||
|
|
Thomas Klauset Aurdal
Advanced Cruncher Joined: Apr 27, 2007 Post Count: 69 Status: Offline Project Badges:
|
I read an interesting article at BBC today about the limitations to bioinformatics when it comes to spot visual objects:
http://news.bbc.co.uk/2/hi/technology/8552471.stm http://solarstormwatch.com/ Can someone answer me on if this applies to WCG and HCC in particular? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I don't believe there is any sort of image processing here, so I'd be surprised if it does.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello Thomas Klauset Aurdal,
I am a bit puzzled by your comment that the BBC article discusses limitations. I read the article and the only limitation they seemed to mention was the ability to analyze so much data. Projects like Solar Stormwatch or Galaxy Zoo (my favorite in this group) are called Distributed Human Projects and use the human ability to make sense out of images. A number of these projects are listed at http://distributedcomputing.info/projects.html A former project, Help Defeat Cancer, used the WCG to analyse images of tissue biopsies to classify cancers with computers. It was quite successful and is now going through more development phases supported by multiple universities. Our current project, Helpm Conquer Cancer, is another image analysis project. We are trying to analyze images of crystal formation using computers. What sort of computational filters work best to characterize the different phases? Lawrence |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thanks for the information!
|
||
|
|
Thomas Klauset Aurdal
Advanced Cruncher Joined: Apr 27, 2007 Post Count: 69 Status: Offline Project Badges:
|
I asked the question because there are some sort of image processing involved in HCC. I were therefore anxious if there might be some barriers to conventional bioinformatics. That conventional bioinformatics might not be sufficient advanced towards solving these WU's. Then thanks for telling me about the experiences with HDC and analyse images of tissue biopsies to classify cancers with computers.
----------------------------------------However, I think there is lots of room for improvement when it comes to bioinformatics. Especially the R&D of Artificial Intelligence (AI). I explored some interesting concepts at DARPA IPTO. Taking for example language processing which is still in it's infancy. http://www.darpa.mil/ipto/ I have full focus on HCC now processing 24/7/365 Cancer is going to kill all of us if we are not doing something about it in time. Therefore I also donate to SENSF and it's OncoSENS project. http://sens.org/index.php?pagename=oncosens [Edit 2 times, last edit by Thomas Klauset Aurdal at Mar 9, 2010 3:39:02 PM] |
||
|
|
|