I've been wondering, with so much of the world of protenomics still in the "unknown" catagory, how does one test a "designer drug" for safety.

When I say designer drug, I basically mean a protein designed to treat a specific disease, such as the current Rosetta docking work units studying AIDS.

And what I mean by safety is... well, your drug (once you find it) is designed to interfere with AIDS (or other disease cells), but with tens of thousands of other proteins streaming through you body, and most of them still of unknown structures, how do you assure that the drug doesn't ALSO interfere with critical or normal bodily functions?

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/

I've been wondering, with so much of the world of protenomics still in the "unknown" catagory, how does one test a "designer drug" for safety.

When I say designer drug, I basically mean a protein designed to treat a specific disease, such as the current Rosetta docking work units studying AIDS.

And what I mean by safety is... well, your drug (once you find it) is designed to interfere with AIDS (or other disease cells), but with tens of thousands of other proteins streaming through you body, and most of them still of unknown structures, how do you assure that the drug doesn't ALSO interfere with critical or normal bodily functions?

Good question. The only definitive way of determining safety is to carry out carefully monitored clinical trials. Predicting the effect of a drug or protein therapeutic on all the complex systems in the body is even harder than predicting protein structure!

---

I've been wondering, with so much of the world of protenomics still in the "unknown" catagory, how does one test a "designer drug" for safety.

...

Good question. The only definitive way of determining safety is to carry out carefully monitored clinical trials. Predicting the effect of a drug or protein therapeutic on all the complex systems in the body is even harder than predicting protein structure!

As discovered by the unlucky volunteers testing a drug recently at Northwick Park Hospital (Greater London, UK) when everyone except the placebo recipients needed to be put on life support machines.

River~~

---

unlucky volunteers testing a drug recently

The volunteers knew what they were getting into, i.e. testing an unknown drug, they do this for money. The rules for proving a drug safety required that the drug be tested on a group not suited to the drug.

It was a drug targetted at people with reduced immune systems, when given to people with healthy immune systems, their systems went into overdrive/beserk, (actually part of the design of the drug - it was designed to boost immune response).

This may have been predicted, but it doesn't matter, the rules required it was tested in the manner it was. The result is that a possibly useful formulation is supressed because the legal process necessary to test it says it is dangerous, a process which is flawed.

Simply...

a + "target group" = cure
a + "control group" = lawyers make a fortune.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

I suppose once you can model a protein from its DNA sequence, then its a logical progression to be able to build a database of the proteins in the body and be able to test new enzymes against all models in the database. It's probably not unrealistic for the near future with the potential CPU power available in the world, especially if we can tap into the consoles.

Find-a-drug built a database of billions of molecules each tested against thousands of different active sites for this reason.

Is there any theoretical reason why you couldn't take the entire DNA sequence for an organism and work out the amino acid sequences it can be transcribed to, and then use Rosetta to create models of each of the proteins? Then you'd have a complete library of all proteins available. Would you end up simulating lots of proteins from unused code which the organism never creates in real life?

---

I suppose once you can model a protein from its DNA sequence, then its a logical progression to be able to build a database of the proteins in the body and be able to test new enzymes against all models in the database. It's probably not unrealistic for the near future with the potential CPU power available in the world, especially if we can tap into the consoles.

Find-a-drug built a database of billions of molecules each tested against thousands of different active sites for this reason.

Is there any theoretical reason why you couldn't take the entire DNA sequence for an organism and work out the amino acid sequences it can be transcribed to, and then use Rosetta to create models of each of the proteins? Then you'd have a complete library of all proteins available. Would you end up simulating lots of proteins from unused code which the organism never creates in real life?

Not a crazy idea at all, but perhaps a bit ahead of its time. We first need to be able to consistently predict structure to very high accuracy, and to dock proteins and small molecules to these predicted structures accurately. We are definitely making progress in all of these areas, so your vision might come to pass sooner rather than later!

---

unlucky volunteers testing a drug recently

The volunteers knew what they were getting into, i.e. testing an unknown drug, they do this for money. ...

agree totally. They were not treated unfairly in any way and (IMHO) didn't deserve any compensation beyond decent care to get well again.

I still say the outcome was still unlucky for them. Lawyers find it hard to grasp, but there *is* such a thing as an unlucky outcome to a known risk. Lawyers will find it even harder to grasp that in such cases there should not be a claim, in my opinion.

R~~

Lawyers will find it even harder to grasp that in such cases there should not be a claim, in my opinion.

Absolutely. I could begin a rant on my dim view of certain factions of the legal profession, but I'm sure the readership here is intelligent enough to make such an endeavour un-necessary.

---

Wave upon wave of demented avengers march cheerfully out of obscurity into the dream.

Wow, timing is everything. News article today, $10m prize for the first to sequence genes of 100 people in 10 days. And if you can do the gene sequencing that quickly, that should be a leg up on performing automated tests to find any adverse side effects on how a designer drug will effect a specific person.

---

Add this signature to your EMail:
Running Microsoft's "System Idle Process" will never help cure cancer, AIDS nor Alzheimer's. But running Rosetta@home just might!
https://boinc.bakerlab.org/rosetta/