| Home |
| Participants |
|
| Content & Focus |
|
| Travel & Lodging |
|
| Presentations & Materials |
|
|
||||||||||||||||||
|
|
|
||||||||||||||||
|
|
|
|
|
|
|
||||||||||||
|
||||||||||||||||||
|
|
|
 |
Biological SystemsBen Shapiro, Alex Bangs, Frank Doyle, Kirk Jordan, Linda Petzold 1a) What are the challenging industry and societal problems to be solved? Funnel problem: 5,000 drugs in, 1 not so good drug out. Now 50,000 in, one not so good out. Pharma continually gets new technology thrown at them, most doesn’t work, have become pessimists. Doing modeling for pharma, but hard to predict outcomes. 99.9% works, 0.1% toxic = very hard to predict. Is a grand challenge problem. Can we ID which people should/shouldn’t have a certain drug? Academic medical research centers do not have resources to do clinical trials. Need cyberinfrastructure to let them do the trials. Can we compile data (who smoked, what age, what prior medical history), to populate clinical trials. Worry about privacy, IP, etc. How to combine small, private pools of specialized data? Outside US, good clinical data now being used to lure good researchers to those countries. In US, data is hidden due to IP, legal issues, bad press, etc. Biological variability >> chemical or engineering variability.
1b) What are the future IT-enabled economic drivers for process and biological systems in the US? [For us, as applied to biological systems.] Enable processes of 1a = big impact. Problem: research community, NIH wants open source, inhibits company ability to function – what do you sell? If companies can’t survive, is bad for research community. Personalized drugs already happening (Herceptin). Proof of principle is out there. If pharma becomes more efficient, can target smaller pools of patients. Can also address rarer diseases. Drug resistant crops. Other non-pharma applications.
2) Where are the gaps in knowledge? What are the problems faced by IT capabilities? [again, for biological systems] We talked more about how to address the knowledge gaps, rather than what they are. For the most part, we basically agree on what they are (communications between experts, multi-scale, multi-physics modeling, education, items from question 1, … ) Bottom line: need resources. This will not happen by volunteer efforts. Will require infra-structure. There have been some attempts at inter-agency efforts. Haven’t seen convergence. Need some kind of body to enforce standards. People want to throw in the kitchen sink when doing modeling. Need to prevent that. Current standard, SBML, does not capture all models. E.g. does not capture any of the models from Entelos. Combat the ‘ODE mafia’. Can’t have one framework for all models. Rather, need good model frameworks for good subclasses. Define sub-classes. Define frameworks. Structure to keep track of model + model assumptions: these models do/don’t fit together. How to transfer data > information > knowledge? Cyber-infrastructure providing the structure to do this. How? How to get people to talk to each other. If on same proposal/center. If meet each other at random (I met my best collaborator playing tennis). How to create an infra-structure that allows people to find common problems / to collaborate? Biology equivalent of math centers? Gateways/other programs at NSF?
3) What are the opportunities for CI? Scale x detail ~ roughly constant (may increase constant a bit by faster computation). Idea: details where you need them. How? Grand challenge problems:
|
|
|||||
|
|||||||