The installer quits straight away unless you use the compatability app and force XP to pretend to be Win98. Once installed, Morpheus works fine without any compatability help.
I'll stick a screenshot somewhere if you *really* want...
Gambling, by definition, is a game directly based on an element of chance. The stock market is based directly on the abilities of companies to generate more money from an inital capital investment. The majority of investment is done by large companies that invest on the basis of significant research.
So horse racing isn't based on the ability of one horse to run faster than the other, or for trainers to train their particular horse to run faster?
Ah fuck it, it's late, I've been at the pub, I probably missed your point:)
Let me start by saying that I like your thinking, but I just cant see it as feasable in any way.
A protein is simply a chain of amino acids, which in turn is simply a chain of bases, of which there are four (well, five, but the fifth isn't usually counted).
A protein is simply a chain of amino acids, that may be modified by glycosolation (actually a major thing to consider), methylation, there may be disulphide bonds added (not really in any predictable position), the chains may be complexed with metals (eg haemoglobin), the protein may only be functional under certain conditions, it may be sequestered by other proteins, it may be free in solution, it may be contained within a lipid membrane...
As for the bases, I presume you mean the DNA bases that encode a protein, i.e. Adenine Threonine Guanine and Cytosine, and the 5th you refer to is Uracil (replaces Threonine in RNA which is transcribed from DNA). To say that Uracil is mostly ignored is quite insulting to a graduate biochemist:) BTW, in real systems some of these bases may be methylated aswell.
Taking your example of A coupling to B, coupling to C, making drug D, you start by reversing the logic. D is made up of [A + B] + C, where A, B and C are sub-units which can be modelled inter-dependently.
I think you need to go back and read again. Its not A to B to C makes D, it's A coupled to B coupled to C makes a functional unit, which interacts with drug D.
Add to that then that A and B, and maybe B and C have seemingly unrelated roles elsewhere. Then in cell type X you have another interacting partner F, which isnt present in your first cell type (for which you took the initial conditions). Introducing drug D to cell type X results in apoptosis (cell death), or worse, necrosis.
How do you decide which (computationally) seemingly benificial reactions, are infact the road to death and general nastyness?
Your previous statement that you only need to model at most 4 bonds (chemical) is also a fallacy. As I tried to allude to earlier, some of the most important bonds in biological systems are hydrogen bonds. Fuzzy bonds if you like. Singly they mean very little, but on mass they are major contributors of the three dimensional structure of proteins. Un my example, A B and C are linked by multiple hydrogen bond interactions. Your initial energy minimization may not even represent the way drug D interacts with these protein subunits. Believe it or not, the lowest possible energy does not always represent the wild type state of a protein.
Given an amino acid sequence we cannot, even now, accurately define the three dimensional structure of a protein. Thats a single protein. Now how on earth do you propose to work out benificial or adverse affects of a single drug on an entire complex system?
You cant even pick initial conditions that are "about right" because everyone is different. Even the conditions inside one individuals body varies incredibly over a few hours. Ever wondered why you can eat peanuts and they will kill another person? This is just the type of randomness you have to deal with. If you like, you may have subunits A B and C, someone else may have A B and G which perform the same tasks, but react entirely differently to Drug D.
The idea of looking at a unified scale is to remove the problems of having things at different scales. Let's take the problem of "is it lipid soluble?" Well, you just don't care. Not at the atomic level. The interaction between entity A and entity B will -emulate- the solubility of entity A, with respect to entity B, even if you never have a specific "check" for that, or even know what to "check" for.
I'm afraid I'd care quite a bit if you came up with a wonder drug that could never reach it's target. A cell isn't just one lipid bag, it has many little bags inside it. If your drug cant reach it's target inside one of these bags it's rather a waste of time, if it reaches and adversly affects a different system, then it's a liability.
But once the body reaches homeostasis
Another quote from a lecture, "the only time reactions reach homeostasis in the body is when you are dead". Really you'ld need a good book on metabolism to understand that, but theres always something driving a reaction.
Animal testing isnt really about finding drugs, it's about finding how harmful and/or effective they are on an entire biological system, before moving onto a human system. As I said, Biology is a knowledge based disiplin. Does it kill the rat? No? Well lets try it on something else. Oh, and you may be suprised just how much we have in common with rats and mice.
Your proposed methods of drug design would be wonderful, if it were not for the fact that there are just too many unknowns for it to be viable (at least in the forseeable future)
I would however encourage you to think about studying biochemistry. With a real understanding of the difficulties behind drug design, maybe one day you may be able to ratianally design an aids cure.
Now, let's move on to the "understanding" of the human body. Excuse me, but who says you HAVE to understand it???
I should hope that if your going to have a go at combating AIDs then you'ld have a damn good understanding of the haman body. After all, as a Virus it does hijack a lot of the native human machinery to do its job.
Its very simple to combat AIDs, whats trickier is combating it and keeping the patient alive.
At any given instant, you can only ever have an interaction between two molecules. (DUH!)
Hardly. Anothing thing about human machinery is it's very rarely just one molecule interacting at a time, rather a number of subunits. If A couples to B which couples to C, and you have drug D, it's interaction with subunit C may cause A to go off and begin a potentially leathal cascade somewhere else.
Now, inorder to do your minimum energy computations you'll need to know the structures of each subunit, and the solvent conditions they are in (probably water, but who knows, they could exist in the lipid bilayers). The PDB is good, but its no where near having all the structures present in the human body. You arnt just dealing with chemical bonds either, there are weaker interactions like Hydrogen bonds to consider. These are trickier to calculate.
And then, maybe your drug (assuming its any good) may never even make it to the target sites. Maybe its not lipid soluble (and so cant get inside the cell), or acid labile (cant be swallowed).
You see, you've tried to over simplify things to the point that your pissing in the wind. In many cases, you really are looking at the ant seriously affecting the elephant, maybe indirectly by affecting the mouse (say bacteria).
That said, rational drug design has worked (IIRC even in designing a few nucliotide analogues used to combat the HIV virus). But the people in sat infront of the computers understand that it's more than a question of crunching numbers. As my Bioinformatics lecturer once said "Biology is a knowledge based disciplin, you work from what youve seen before. Physics/computing is ultimately easier, because you can work from first principles". We just dont know anywhere near enough of Biologies first principles to even think of moving completely to rational drug design.
We had a little discussion on alt.comp.peripherals.videocards.nvidia with a guy doing CAD. Turns out this feature is quite useful to him, as it lets him quickly see *all* of his design. Apparently the software alone can manage something similar, but it's a lot more hassle than Now you see everything, Now you dont.
If it's quake3 youre developing maps for, start the map with/devmap Bring down the console again, and type/r_showtris 1 (at least I think thats it).
Overlays a wireframe of whats actually being drawn IIRC. Of course, the r_showtris command wont work in a real game...
Slashdot Mirror
Actually, it does work. With XP home at least.
The installer quits straight away unless you use the compatability app and force XP to pretend to be Win98. Once installed, Morpheus works fine without any compatability help.
I'll stick a screenshot somewhere if you *really* want...
Gambling, by definition, is a game directly based on an element of chance. The stock market is based directly on the abilities of companies to generate more money from an inital capital investment. The majority of investment is done by large companies that invest on the basis of significant research.
:)
So horse racing isn't based on the ability of one horse to run faster than the other, or for trainers to train their particular horse to run faster?
Ah fuck it, it's late, I've been at the pub, I probably missed your point
A protein is simply a chain of amino acids, which in turn is simply a chain of bases, of which there are four (well, five, but the fifth isn't usually counted).
A protein is simply a chain of amino acids, that may be modified by glycosolation (actually a major thing to consider), methylation, there may be disulphide bonds added (not really in any predictable position), the chains may be complexed with metals (eg haemoglobin), the protein may only be functional under certain conditions, it may be sequestered by other proteins, it may be free in solution, it may be contained within a lipid membrane...
As for the bases, I presume you mean the DNA bases that encode a protein, i.e. Adenine Threonine Guanine and Cytosine, and the 5th you refer to is Uracil (replaces Threonine in RNA which is transcribed from DNA). To say that Uracil is mostly ignored is quite insulting to a graduate biochemist :) BTW, in real systems some of these bases may be methylated aswell.
Taking your example of A coupling to B, coupling to C, making drug D, you start by reversing the logic. D is made up of [A + B] + C, where A, B and C are sub-units which can be modelled inter-dependently.
I think you need to go back and read again. Its not A to B to C makes D, it's A coupled to B coupled to C makes a functional unit, which interacts with drug D.
Add to that then that A and B, and maybe B and C have seemingly unrelated roles elsewhere. Then in cell type X you have another interacting partner F, which isnt present in your first cell type (for which you took the initial conditions). Introducing drug D to cell type X results in apoptosis (cell death), or worse, necrosis.
How do you decide which (computationally) seemingly benificial reactions, are infact the road to death and general nastyness?
Your previous statement that you only need to model at most 4 bonds (chemical) is also a fallacy. As I tried to allude to earlier, some of the most important bonds in biological systems are hydrogen bonds. Fuzzy bonds if you like. Singly they mean very little, but on mass they are major contributors of the three dimensional structure of proteins. Un my example, A B and C are linked by multiple hydrogen bond interactions. Your initial energy minimization may not even represent the way drug D interacts with these protein subunits. Believe it or not, the lowest possible energy does not always represent the wild type state of a protein.
Given an amino acid sequence we cannot, even now, accurately define the three dimensional structure of a protein. Thats a single protein. Now how on earth do you propose to work out benificial or adverse affects of a single drug on an entire complex system?
You cant even pick initial conditions that are "about right" because everyone is different. Even the conditions inside one individuals body varies incredibly over a few hours. Ever wondered why you can eat peanuts and they will kill another person? This is just the type of randomness you have to deal with. If you like, you may have subunits A B and C, someone else may have A B and G which perform the same tasks, but react entirely differently to Drug D.
The idea of looking at a unified scale is to remove the problems of having things at different scales. Let's take the problem of "is it lipid soluble?" Well, you just don't care. Not at the atomic level. The interaction between entity A and entity B will -emulate- the solubility of entity A, with respect to entity B, even if you never have a specific "check" for that, or even know what to "check" for.
I'm afraid I'd care quite a bit if you came up with a wonder drug that could never reach it's target. A cell isn't just one lipid bag, it has many little bags inside it. If your drug cant reach it's target inside one of these bags it's rather a waste of time, if it reaches and adversly affects a different system, then it's a liability.
But once the body reaches homeostasis
Another quote from a lecture, "the only time reactions reach homeostasis in the body is when you are dead". Really you'ld need a good book on metabolism to understand that, but theres always something driving a reaction.
Animal testing isnt really about finding drugs, it's about finding how harmful and/or effective they are on an entire biological system, before moving onto a human system. As I said, Biology is a knowledge based disiplin. Does it kill the rat? No? Well lets try it on something else. Oh, and you may be suprised just how much we have in common with rats and mice.
Your proposed methods of drug design would be wonderful, if it were not for the fact that there are just too many unknowns for it to be viable (at least in the forseeable future)
I would however encourage you to think about studying biochemistry. With a real understanding of the difficulties behind drug design, maybe one day you may be able to ratianally design an aids cure.
Maybe.
Now, let's move on to the "understanding" of the human body. Excuse me, but who says you HAVE to understand it???
I should hope that if your going to have a go at combating AIDs then you'ld have a damn good understanding of the haman body. After all, as a Virus it does hijack a lot of the native human machinery to do its job.
Its very simple to combat AIDs, whats trickier is combating it and keeping the patient alive.
At any given instant, you can only ever have an interaction between two molecules. (DUH!)
Hardly. Anothing thing about human machinery is it's very rarely just one molecule interacting at a time, rather a number of subunits. If A couples to B which couples to C, and you have drug D, it's interaction with subunit C may cause A to go off and begin a potentially leathal cascade somewhere else.
Now, inorder to do your minimum energy computations you'll need to know the structures of each subunit, and the solvent conditions they are in (probably water, but who knows, they could exist in the lipid bilayers). The PDB is good, but its no where near having all the structures present in the human body. You arnt just dealing with chemical bonds either, there are weaker interactions like Hydrogen bonds to consider. These are trickier to calculate.
And then, maybe your drug (assuming its any good) may never even make it to the target sites. Maybe its not lipid soluble (and so cant get inside the cell), or acid labile (cant be swallowed).
You see, you've tried to over simplify things to the point that your pissing in the wind. In many cases, you really are looking at the ant seriously affecting the elephant, maybe indirectly by affecting the mouse (say bacteria).
That said, rational drug design has worked (IIRC even in designing a few nucliotide analogues used to combat the HIV virus). But the people in sat infront of the computers understand that it's more than a question of crunching numbers. As my Bioinformatics lecturer once said "Biology is a knowledge based disciplin, you work from what youve seen before. Physics/computing is ultimately easier, because you can work from first principles". We just dont know anywhere near enough of Biologies first principles to even think of moving completely to rational drug design.
We had a little discussion on alt.comp.peripherals.videocards.nvidia with a guy doing CAD. Turns out this feature is quite useful to him, as it lets him quickly see *all* of his design. Apparently the software alone can manage something similar, but it's a lot more hassle than Now you see everything, Now you dont.
If it's quake3 youre developing maps for, start the map with /devmap Bring down the console again, and type /r_showtris 1 (at least I think thats it).
Overlays a wireframe of whats actually being drawn IIRC. Of course, the r_showtris command wont work in a real game...