An Industrial Perspective on Scientific Computing
Dan Sulzbach, Director
Scientific Computing
Genentech, Inc.
sulzbach@gene.com

Genentech is a research driven Biotechnology company with approximately 2,500 employees. The Scientific Computing Department at Genentech supports the computing and information needs of corporate R&D activities. From my "Biotechnology industry point of view," scientific computing encompasses a much broader range of activities and requirements than is sometimes assumed.

"Scientific computing" at Genentech and other Biotechnology and Pharmaceutical companies includes computing activities from the very initial discovery of potential therapeutic agents to ultimate regulatory review and approval. This encompasses the computing and information needs of research scientists, assay developers and assay labs, preclinical testing, clinical testing, and regulatory submission. Although the specific details will vary from industry to industry, the general process of initial discovery, followed by testing and refinement, leading to ultimate development of a product that adheres to industry standards or regulations is common to most industries.

Because of the range of scientific activities at Genentech, scientific computing encompasses many things. Not all of these are typically identified as being part of "scientific computing." But, they are. During the past year, I and my staff have been involved in many activities, including the following:

• computational analysis and simulation (e.g., Computational Chemistry, Molecular Modeling, and structure refinement),
• designing and developing more efficient and user-friendly data management systems,
• facilitating access to and interpretation of electronic information resources,
• Bioinformatics,
• computer security,
• electronic documentation and document management,
• lab automation and robotics, and
• designing and developing Graphical User Interfaces to applications.

Future scientific computing activities at Genentech will be shaped by the following technology trends:

• increasingly powerful computers that are both affordable and supportable,
• faster and more widespread networks and their concomitant resources,
• Biological data bases that are growing rapidly in size and complexity,
• improving data storage and management technology, and
• digital imaging technology.

How well are students being prepared to work in this type of environment? In the last year we have hired 8 new full-time staff. Only one of these was hired directly out of school. She was in a graduate program in Computer Science. The others all had prior work experience. Without this experience, they would not have had the expertise and skills we needed.

I end with some personal observations and opinions. To wit:

• "Computational Science" is only one part of Scientific Computing.
• Data management is currently an underemphasized aspect of Scientific Computing.
• Universities will probably never be able to provide some of the information obtained from practical experience (e.g., "industrial strength" experience with commercial software for large document management).
• Computer security is becoming critically important to Scientific Computing in an industrial setting.