Print

Inside Informatics 12.4


History of Chemical Information Progress

CambridgeSoft's Dr. Jim Dill takes a look back at how far we have come. 


James D. Dill, Ph.D.

 

IN AUGUST, CambridgeSoft held its Users' Group Meeting. The meeting organizers invited me to present a sweeping general overview of trends in the chemical information field. I protested that I didn't have one, that my view of the field has been long, but not very broad. Twenty-five years ago, I set out to solve a small set of problems, and since then I've been re-solving the same few problems over and over: how to represent a molecule in the computer, how to deal with a lot of molecules, how to handle a chemical reaction, how to present chemical data to a user. In this presentation, I'll compare how we approached some of these problems in the early days of my career versus how we look at them today.

The Early Days

First, let me provide some background. The year is 1978; I am finishing up a postdoc, looking for a job, when I hear news my old friend Todd Wipke and two cohorts are starting a company. I sign up, drive across the country, show up for work one Monday morning in Hayward, California, and knock on the door. There is no answer. I wait. Hours later, up comes a bushy beard saying "Hi, I'm Steve, you must be the new guy." He opens the door, and Molecular Design Ltd. is open for business.

There were four of us. Steve Peacock was the hardware whiz, knower of all jargon and expert on auto repair. Todd Wipke was a newly-tenured professor at Santa Cruz, bringer of algorithms and Fortran code. Stuart Marson was more entrepreneur than chemist, a great salesman, and one-of-a-kind personality (that is, unless you know anyone else named Stewart who started a chemical software company and has a good head of hair!).

Our plan was to assemble an arsenal of computational chemistry programs into a system for designing molecules to order. We would take a customer's data, run it through our machine, and out would come potent drug candidatesÑthat was the plan. We began collecting programs and soon ran across the first problem to be solved: how to represent a molecule in some portable format which could be transferred between programs. A series of evening e-mails ensued, resulting in what Todd called the Standard Molecular Utility (SMUT) File, fortunately later renamed the Molfile before it became a world-wide standard.

We agonized over whether the smutfile should be readable ASCII or compact binary, and opted for the former even though it consumed precious disk space. So how do we represent molecules today? .....

To read the full article, Please click here to login into CambridgeSoft.com 

 
All trademarks are the property of their respective holders.
©2022Copyright 1998- PerkinElmer Inc. All Rights Reserved. Privacy Statement