Recognition of ill-defined signals in nucleic acid sequences

A set of programs has been developed for the definition andhandling of nucleic acid sequence consensus information. Thesequences of known genetic control signals are combined in amatrix. The origins and positions of the signals are recorded.Old matrices can be updated dynamically: new signals are includedand obsolete ones deleted. Matrices of several different typesare computed optionally. Several of these matrices can be combinedto find possible new signals. The use of matrices allows theexact quantification of signal qualities. The described programsare part of a program library named GENEXPERT. Application examplesgiven are the search for tRNA genes and the search for promotersin the bacteriophage genome. Received on July 22, 1987; accepted on October 5, 1987