We have developed an automated system for the design of laboratory experiments in molecular biology. The system uses a planning method known as skeletal plan refinement that attempts to emulate the human cognitive task of experiment design. This paper describes the theory, history, and implementation of the design system and illustrates its function in the domain of DNA cloning experiments.

We have built a knowledge-based genetic engineering simulation system-- GENESIS-- capable of representing both domain-specific and general knowledge. Information is stored within a hierarchically-organized framework composed of structures called units. A series of sophisticated editors enables no-computer specialist molecular geneticists to construct a knowledge base through direct interaction with the computer. Three types of knowledge specific to the domain of molecular genetics, MAPS, sequences and RULES are discussed in detail with examples.

The MAXAMIZE advisory system determines from user-provided restriction maps an optimal strategy to do nucleotide sequencing by methods involving end-labeled fragments. The maps may be either simple linear restriction maps of fragments or complex circular maps including restriction sites of a vector. The whole system is interactive and is written in the Genetic English language provided by the GENESIS System, a molecular genetics knowledge representation and manipulation package. In addition, MAXAMIZE provides bookkeeping facilities for sequencing and offers advise on how to verify the newly obtained sequence data.