Splicing UNIX into a genome mapping laboratory

The Whitehead Institute/MIT Center for Genome Research is responsible for a number of large genome mapping efforts, the scale of which create problems of data and workflow management that dictate reliance on computer support. Two years ago, when we started to design the informatics support for the laboratory, we realized that the fluid and ever-changing nature of the experimental protocols precluded any effort to create a single monolithic piece of software. Instead we designed a system that relied on multiple distributed data analysis and processing tools knit together by a centralized database. The obvious choice of operating systems was UNIX. In order to make this choice palatable to the laboratory biologists—who rightly consider it their job to do experiments rather than to interact with computers, and who have come to expect all software to be as intuitive and responsive as the Apple Macintoshes on their desks—we designed a system that runs automatically and essentially invisibly. Whenever it is necessary for the informatics system to interact with a member of the laboratory we have carefully chosen a user interface paradigm that best balances the user’s expectations against the system’s capabilities. When possible we have chosen to adapt familiar software to our user interface needs rather than to write user interfaces from scratch. We’ve managed to hide the power of UNIX behind the innocuous personal computer-based front ends our users know and love, using techniques that should be applicable in other environments as well.