Log In Join Now
Forums  |  Help  |  Settings  |  Download
Research: FightAIDS@Home: Project FAQs Collapse   Expand  

FightAIDS@Home is a project of the Olson laboratory at The Scripps Research Institute that uses volunteer computing technology which allows you to contribute your device's idle resources to accelerate research into new drug therapies for HIV, the virus that causes AIDS.

All you need to do to join FightAIDS@Home is download and install the free software. Once that has been done, your device is then automatically put to work, but you can also continue using it as normal. Click here to get started.

At any one time, the project uses one of two software docking tools to automatically download small pieces of work to your device and performs calculations that model how drugs interact with various HIV virus mutations. After your device processes the information, the results are sent back to World Community Grid and then sent on to The Scripps Research Institute where they are analyzed by the Scripps research team. The process takes an enormous amount of computing time, which is why World Community Grid needs you (and your friends!) to participate in FightAIDS@Home project.

Your device will contribute to whatever projects you choose; however, only certain projects will be available for mobile devices. You can select from the projects currently active at World Community Grid by visiting the My Projects page. There you can view all available projects, and choose those in which you want to participate.

You may find the latest status on the FightAIDS@Home Project here.

AutoDock and Vina are automated docking software tools. They are designed to predict how a small molecule, such as a substrate or drug candidate, binds to a receptor molecule of known 3D structure. In the context of this project, these docking tools are being used to find potential drug compounds which may inhibit the HIV-1 protease (a protein which encourages and controls the progression of the virus).

The two software programs use different algorithms, each of which may provide better results depending on the types of molecules being docked. The FightAIDS@Home project uses both software tools in its calculations: the Scripps researchers determine ahead of time which software package is more suited to the particular task at hand, and the selected software for those work units then runs on World Community Grid. The project may therefore switch back and forth between the two software packages depending on its needs. As a contributor to the FightAIDS@Home project, you may notice either of those software packages being run for this project, each of which has a unique screen saver (see below for details on both screen savers).

Click on the on your agent application window in the lower right hand corner. You then will see a graphics window similar to the following AutoDock screen saver image:

What is the white arrow, helix and loopy structure?
Ribbon diagrams are simplified drawings of proteins that make it easier for scientists to view and understand what is shape is. The three-dimensional "skeleton" of HIV-1 protease is shown as a white ribbon diagram on the screen and is magnified about 10,000,000 times.

In this panel, you can see the shape that the particular sequence of amino acids in HIV-1 protease makes in three dimensions. For clarity, we are not showing the details of all of the atoms in the protein molecule, just the backbone. Remember, all proteins, including HIV-1 protease, are made up of strings of amino acids, linked like beads on a string. There are twenty different naturally-occurring amino acids, and you can think of them as different kinds of building blocks. These strings of amino acids have parts that like to stick to others while repelling others. The different parts of the protein's amino acid chain clump together into characteristic three-dimensional shapes.

The search algorithm used in AutoDock is not just looking at one possible solution of one candidate drug molecule (ligand) but is actually evaluating many possible solutions at once. The spheres show places where the best drug molecule to HIV-1 protease dockings have been calculated and the color shows how good they are.

AutoDock is trying to find the best way that the current ligand, the one your agent has downloaded, can fit together with the target HIV-1 protease. You can think of the ideal drug we are trying to find as a "key," and the HIV-1 protease as a "lock." Unlike keys in the real world, however, many drug molecules bend to change shape. In this respect, molecules are like a dancer's body; the same body is able to adopt many different poses and shapes. Unfortunately, we do not know what shape a candidate drug will adopt until we try millions of different possibilities and then select the best one.

To find the best fit, we are using an algorithm. An algorithm is just a recipe, a list of ingredients and instructions on how to do or make something. We are actually applying the principles of evolution in our search algorithm to find the best way that our candidate drug molecule would best fit together with the target, HIV-1 protease. Like evolution in the real world, we have a "population" of possible solutions to the problem.

This is what you are seeing when you look at the different colored spheres dotted around the white ribbon diagram. The colors correspond to the same colors of the crosses in panel B. Those representing more negative energy are considered better dockings. AutoDock uses a representation for each of these ligand dockings that says where the ligand's center is, what its orientation is, and what shape it has currently adopted. AutoDock applies genetic operations on the representations of random pairs of ligand shapes to generate two new representations and hence potentially better solutions. You can see how well AutoDock is doing by looking at the graph in panel C.

We see here the energy breakdown for each candidate ligand docking of the current population of possible solutions. The total energy of a ligand binding to the HIV-1 protease consists of an electrostatic energy component and a non-bonded energy component. The electrostatic energy measures how many like-charges and unlike-charges are interacting between the ligand and the protease. The non-bonded energy measures non-electrostatic attraction between the two.

You can see electrostatic forces in action if you rub a balloon on a dry wooly sweater, and then gently place the balloon against a wall: It sticks! This is because all objects are made of atoms. Each atom has an equal number of electrons and protons. Electrons have a negative charge, while protons have a positive charge. These charges balance one another exactly to make objects neutral, or uncharged. When we rub the balloon against a sweater, the friction causes electrons to be rubbed off the sweater and onto the balloon. The balloon becomes charged with static electricity, and it now has more electrons than protons, so it is negatively charged; the wall is more positively charged than the balloon so the balloon sticks.
If you were to rub a second balloon on your sweater, and hang the two balloons from a string, you would see the two balloons repel one another.

Non-bonded energy arises because atoms are "sticky" when they get close to one another. The amount of "stickiness" depends on the two atoms that are interacting. However, atoms repel one another when they are pushed too close together. Between two touching molecules, there are many of these non-bonded interactions. They are called "non-bonded" because these interactions are not permanent like chemical bonds.

We see here the best docking energy in the current population, plotted over the course of the current docking, shown as a green solid line. The red-dotted line shows the same kind of graph, but for the best docking achieved so far. As the current docking proceeds, at the end of every generation, the green graph gets updated.

The vertical axis shows the best energy. The more negative the energy, the better, i.e. the more precisely we predict this particular ligand will bind to the protease. You can see times when the energy is not changing (the horizontal lines in the graph) and times when the energy dropped (the vertical lines) when AutoDock has found a better solution than the previous generation.

The Current Progress Bar shows how much of the current work unit has been completed. The work units are specified by the researchers at The Scripps Research Institute and transmitted via the servers at World Community Grid to your machine. Each work unit has just one candidate drug molecule, out of a vast library of candidate drug molecules we are virtually screening. The software running under the grid agent on your device is called AutoDock, and it tries to determine the best way the current ligand fits into the target HIV- 1 Protease. When the work unit is finished, the best results are sent back to Scripps via World Community Grid for further analysis, to find the best candidate protease inhibitors for further testing in the laboratory.

Here is a video of the FightAIDS@Home project graphics when the Vina software is being used:

The right portion of the screen saver shows both the target and drug candidate molecules, depicted as a collection of small spheres that represent the atoms of each molecule. These are the specific molecules that your device is currently working on. The left portion of the screen saver shows a view of the Earth and a red AIDS ribbon representing the worldwide fight against AIDS.

SCRIPPS is the logo for "The Scripps Research Institute" in La Jolla, California, USA, which is the home of the research team behind the FightAIDS@Home project.

The progress bar towards the bottom of the screen saver represents approximately how much of the current task your device has processed. When it reaches 100%, the computation is complete and the results will then be sent back to the World Community Grid servers, where they will be packaged and delivered to the FightAIDS@Home researchers.

The small spheres represent the atoms in both the target molecule and candidate molecule currently being processed by your device.