Okay, halfcard.

Here is the all-text version of http://fightaidsathome.scripps.edu/news/vol1.html

FightAIDS@Home News
January 19th, 2006 Volume I

More than a quadrillion energy evaluations...
We have reached the half-way point in screening thousands of diverse drugs against hundreds of HIV protease variants.

As most of you are aware, the overall goal of the FightAIDS@Home computational effort is to help find promising new leads for HIV therapies. We are focused on a very important aspect of HIV drug therapy, the evolution of drug resistance and how to develop new drugs that will be more robust against this continual process. In order to approach this problem, we are undertaking a huge computational endeavor: simulating the the co-evolution of virus and drug to an atomic level of detail, and looking for the best drugs in this “arms race.” In computational terms, this means computing the effectiveness of all possible drugs against all possible mutations of the virus. This is a problem too large even for all of the world’s computational resources. In practical terms we are looking at very large libraries of chemical compounds and evaluating their interactions against large panels of known and potential mutations of an important protein of the virus, HIV protease, which is an essential component of the viral life cycle. Even this computation is enormous, and must be broken down into stages.

We are currently running the first stage of these computations, called Stage 1a. This involves screening a large library of chemical compounds available from the National Institutes of Health against a representative panel of HIV protease mutants. Right now, we are looking at a subset (almost 1,900 compounds) of this library that have been chosen to represent the chemical diversity of the complete library (with over 200,000 compounds). We are using our AutoDock program to dock each of these 1,900 compounds against a panel of 270 variations of the HIV protease structure: that’s 513,000 experiments. Every time we want to predict how a small chemical compound might fit into the large HIV protease structure in one of these experiments, we have to compute the energy of interaction of 10 million different ways of fitting the compound into different places on the protein target. Because each docking calculation has so many variable s, we actually do the dockings 100 times for each candidate-drug/protease combination. Moreover, to guarantee the fidelity of the results, we send the same candidate-drug/protease combination to 5 different computers. So this stage of our effort represents over half a million different drug/target pairs, with over 256 million dockings for each pair. That’s over two and a half quadrillion energy evaluations! (a quadrillion is a thousand raised to the 5 th power, or 1,000,000,000,000,000). Remember, each energy evaluation considers all the non-bonded and electrostatic interactions between some 50-100 atoms in the compound against all the atoms in HIV protease, nearly 2,000 atoms.

We have now been running FightAIDS@Home on World Community Grid for almost 2 months, and have made great progress on a number of fronts. Firstly, we are gratified to see the growth in membership of World Community Grid and participation in FightAIDS@Home. Over 26,000 new members with over 45,000 devices have joined since we launched and that boost in computational power has meant a dramatic increase in the rate of results that we have received. Together, we have now crunched our way through 135 of the 270 HIV protease variants, in each case screening against all 1,900 library compounds. Thus we anticipate that Stage 1a will be finished in another two months. As a control, we included the clinically approved HIV protease inhibitors, and have seen that the docking results return the experimentally observed structures and relative energies of binding. We are now evaluating the returned results and developing new tools to help us mine this mind-boggling amount of data.

Stage 1b i s a follow up where we will dock all 200,000 compounds in the NIH library against the ‘wild-type’ HIV protease. This will tell us if the best compounds found in the smaller ‘diversity set’ missed any of the best compounds we find in the second stage. We will be able to confirm this through a process called “similarity searching”. If new classes of compounds are discovered in this follow up stage ( i.e. 1b), they will be added to our list of “lead” compounds. Stage 1b, which will start some time in March 2006, will require 200,000 candidate-drug/protein dockings, and thus will take a little less than half the time of stage 1a, which had 513,000 pairs of molecules to dock.

Finding promising leads through FightAIDS@Home is one component in our larger HIV research project, which involves a number of other laboratories doing experimental work in x-ray crystallography, molecular biology, synthetic chemistry and cell biology. The overall project has also been making significant progress, in part based on our earlier computational results. For those of you interested in the scientific details of some of this work, we have published three new papers in highly regarded scientific journals. One describes the modeling and structural work on earlier discovered lead compounds against both wild type and drug resistant mutants of HIV protease. Additionally, a second paper will be coming out shortly describing new lead compounds that have been selected for by HIV protease itself, using a novel synthetic approach termed “Click Chemistry,” developed in the laboratory of our collaborator and Nobel Laureate, Prof. Barry Sharpless.

Prof. Arthur J. Olson
Dr. Garrett M. Morris
Dr. William Lindstrom
Alexandre Gillet

References

1. Brik, A., et al. , 1,2,3-triazole as a peptide surrogate in the rapid synthesis of HIV-1 protease inhibitors. Chembiochem , 2005. 6 (7): p. 1167-9.

2. Heaslet, H., et al. , Structural Insights into the Mechanisms of Drug Resistance in HIV-1 Protease NL4-3. J Mol Biol, 2005.

3. Whiting et al. Inhibitors of HIV-1 Protease through In Situ Click Chemistry. Angewandte Chemie , 2005.