The ability to delineate (at least in theory) all the proteins encoded in the human genome and all of those encoded by the genomes of major human parasites has given us an unprecedented opportunity to address the causes and treatment of every major human disease. However, the vast increase in biological knowledge that has resulted from the last decade of genomic DNA sequencing has led us to a a crisis in bioinformatics. This crisis is two-fold: analysis of data and planning of experiments. Most of the scientific resources being expended in bioinformatics are being spent on data analysis tools. While these are essential, we should not neglect the opportunity to accelerate the progress of actual experimental biology. All modern experimental molecular biology (and, increasingly, structural biology) depends upon the availability of plasmid clones to address specific scientific questions. Although software facilitating DNA manipulations exists, few programs advise users of optimal strategies and none automate the process of clone generation. Genomics initiatives identify proteins at the genome level and demand the generation of hundreds of expression clones for recombinant protein production in exogenous hosts such as E. coli. Establishment of libraries of expression clones requires automation and optimization as well as effective means of data storage, archiving, annotation and query. To address these needs, as well as to facilitate routine DNA manipulations in virtually any molecular biology laboratory, we propose (1) to test and build a task centered virtual cloning expert system that serves as a knowledge base for DNA manipulations, and (2) to test and build an information automaton for the construction of expression clone libraries in support of structural genomics initiatives and other high throughput experiments.
