High computational power reduces time to discovery in protein-fold research at UCLA

Los Angeles 14 August 1998 By the year 2050, scientists will have unveiled the complete DNA genome as well as the 3D structure of each identifiable protein. Diseases caused by mistakes in proteins will be tracked down to observe how the mutations occur. Drugs and therapies will be developed to repair the initial deviation in the errant protein molecules. This is the picture offered by Dr. Duilio Cascio, Research Faculty Member of the Molecular Biology Institute at UCLA, the University of California in Los Angeles. Here, the biologists use custom-designed software to perform compute intensive simulations and molecular modelling on Digital Unix based Alpha workstations. The cost-effective but powerful computer infrastructure enables the scientific team to produce solutions in the shortest time possible.

Advertisement

By the year 2050, scientists will have unveiled the complete DNA genome as well as the 3D structure of each identifiable protein. Diseases caused by mistakes in proteins will be tracked down to observe how the mutations occur. Drugs and therapies will be developed to repair the initial deviation in the errant protein molecules. This is the picture offered by Dr. Duilio Cascio, Research Faculty Member of the Molecular Biology Institute at UCLA, the University of California in Los Angeles. Here, the biologists use custom-designed software to perform compute intensive simulations and molecular modelling on Digital Unix based Alpha workstations. The cost-effective but powerful computer infrastructure enables the scientific team to produce solutions in the shortest time possible.

The secret of all life on this planet lies in the genetic code. This DNA structure was discovered in 1953 by Francis Crick and James Watson. If scientists learn to understand the precise function of both the DNA and all related protein molecules, human illnesses can be traced to their molecular source and possibly be cured. The Molecular Biology Institute is actively engaged in this research for the purpose of which the biologists apply the newest computing and instrumentation technology. The previous 32-bit architecture has been replaced by Alpha's 64-bit capacity in order to tackle an entire simulation at once instead of having to divide the problem into smaller entities.

Each biologist at the institute has an Alpha workstation at his desk side. A high speed Asynchronous Transfer Mode (ATM) network interconnects all the workstations which are equipped with Digital's "Powerstorm" graphics cards. As a result, the simulations as well as the modelling can be conducted on the same machine. For the macro-molecular modelling which is applied on proteins, the specialized "O" graphics software was developed at the Uppsala University in Sweden. In addition, the institute employs a software based failover system, referred to as DECsafe Available Server, to monitor the permanent functionality of the Alpha systems. If a failure is tracked in one system, DECsafe shifts the processing load within seconds to another one without any loss of data.

Platform Computing Corporation's LSF software has been installed to obtain the fastest possible run time on submitted computing jobs. In this way, the institute avoids a waste of CPU cycles and even detects usage trends as well as needs for new computer equipment. Since off-the-shelf software packages are hardly available in the field of protein-fold research, the scientists have written specific and customized programmes. Researchers from all over the world can access these tools via the Internet. At the institute, three servers are operating, two of which are used around-the-clock to execute an average of 300 intense computational jobs per day for submitters worldwide.

Protein folds consist of three-dimensional angles. Amino acid molecules are combining with these angles to form larger protein molecules. Diverse amino acid sequences generate different folds, which determine the final functions of the created protein molecules. In this way, researchers at the institute have sequenced all 4000 proteins generated by pyrobaculum aerophilum, a primitive undersea organism. Since the proteins in this organism share the same mechanisms as those in the human one, this work is fundamental to discover both the structure and function of between 60.000 to 200.000 proteins that humans produce to survive. Dr. Cascio therefore continues to stress the importance of sufficient computing power to free mankind from disease. You can find more information on this kind of scientific research in the VMW article Supercomputer database reveals hidden relationship of proteins with entirely different functions.


Leslie Versweyveld

[Medical IT News][Calendar][Virtual Medical Worlds Community][News on Advanced IT]