I wish I could have posted this question earlier, but, if anyone in the know is reading this, is there a newsgroup for AIDS or a site like stackoverflow.com for medicine?
It hit me after reading this article that perhaps researchers could take a step back and try to look at AIDS as a really hard math problem. That way they could take a higher level view and be able to apply mathematical transformations without getting lost in all the low-level biology and chemistry.
So for example, here are some rules off the top of my head, if we use A for AIDS, C for cells, I for immune system cells, T for time, P for protein, D for drug, R for the rate of mutation, etc:
1. Once A1 is within C1, it can’t be removed
2. C1 expires after some duration T1
3. A1 is invisible to I1 until P1 binds to A1
4. P1 can be delivered to the body by D1
5. P1 can be produced by the body in response to D2
6. A1 mutates into A2 at rate R1
...
After applying a solver: Dn triggers immune system response that create P1, P2..Pn which bind to A1, A2…An and keep An from leaving Cn for T years.
I don’t know what the form of these rules would be or how many there would be. But if they were rearranged into a big matrix in a language like Prolog, then computer simulations could be run that might reveal how to administer some drug that triggers a response that creates something that interacts with the AIDS virus and weakens it or makes it visible to the immune system. Maybe that drug could be a lot cheaper than the current antiviral drugs. Or maybe it could even be a one-time dose that causes the body to make that certain something from then on, working at least until the virus mutates.
I did a quick google search of “protein that binds to AIDS virus” and found these:
They look like mumbo jumbo to me but some things jumped out, like this quote from the second link:
“Native dodecameric SP-D bound to HIV gp120 more strongly than native trimeric SP-D. Since one common polymorphic form of SP-D is predominantly expressed as trimers and associated with lower blood levels, these individuals may have less effective innate defence [sic] against HIV.”
This is probably old news to experts but it was certainly news to me that things like this are already present in the body and come in different forms in different people and have a lot to do with an individual’s susceptibility. I think the web could be used to bring non-medical folks “up to speed” on certain key insights like this. Big data APIs like IBM’s Watson could find the hidden relationships by looking not just at links but at the semantics of the pages.
I’m just saying that everyone could use a little help from outside their chosen field, and building these rulesets and crunching them in big distributed computer networks is really easy for computer scientists (and hackers, honestly). We are used to hammering on a problem until the solution is found, and we are compelled to keep after it, especially if the solution is elusive. If we had a standard notation to go between biology and mathematics, then it would open up a lot of possibilities for kind of “open sourcing” medical research.
I realize this is probably already a work in progress somewhere, but if this isn’t a mainstream thing like SETI@home, then it doesn’t quite exist yet, you know?
It hit me after reading this article that perhaps researchers could take a step back and try to look at AIDS as a really hard math problem. That way they could take a higher level view and be able to apply mathematical transformations without getting lost in all the low-level biology and chemistry.
So for example, here are some rules off the top of my head, if we use A for AIDS, C for cells, I for immune system cells, T for time, P for protein, D for drug, R for the rate of mutation, etc:
1. Once A1 is within C1, it can’t be removed
2. C1 expires after some duration T1
3. A1 is invisible to I1 until P1 binds to A1
4. P1 can be delivered to the body by D1
5. P1 can be produced by the body in response to D2
6. A1 mutates into A2 at rate R1
...
After applying a solver: Dn triggers immune system response that create P1, P2..Pn which bind to A1, A2…An and keep An from leaving Cn for T years.
I don’t know what the form of these rules would be or how many there would be. But if they were rearranged into a big matrix in a language like Prolog, then computer simulations could be run that might reveal how to administer some drug that triggers a response that creates something that interacts with the AIDS virus and weakens it or makes it visible to the immune system. Maybe that drug could be a lot cheaper than the current antiviral drugs. Or maybe it could even be a one-time dose that causes the body to make that certain something from then on, working at least until the virus mutates.
I did a quick google search of “protein that binds to AIDS virus” and found these:
http://en.wikipedia.org/wiki/Envelope_glycoprotein_GP120
http://vir.sgmjournals.org/content/86/11/3097.full
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC191382/
http://www.pnas.org/content/111/26/E2676.abstract
They look like mumbo jumbo to me but some things jumped out, like this quote from the second link:
“Native dodecameric SP-D bound to HIV gp120 more strongly than native trimeric SP-D. Since one common polymorphic form of SP-D is predominantly expressed as trimers and associated with lower blood levels, these individuals may have less effective innate defence [sic] against HIV.”
This is probably old news to experts but it was certainly news to me that things like this are already present in the body and come in different forms in different people and have a lot to do with an individual’s susceptibility. I think the web could be used to bring non-medical folks “up to speed” on certain key insights like this. Big data APIs like IBM’s Watson could find the hidden relationships by looking not just at links but at the semantics of the pages.
I’m just saying that everyone could use a little help from outside their chosen field, and building these rulesets and crunching them in big distributed computer networks is really easy for computer scientists (and hackers, honestly). We are used to hammering on a problem until the solution is found, and we are compelled to keep after it, especially if the solution is elusive. If we had a standard notation to go between biology and mathematics, then it would open up a lot of possibilities for kind of “open sourcing” medical research.
I realize this is probably already a work in progress somewhere, but if this isn’t a mainstream thing like SETI@home, then it doesn’t quite exist yet, you know?