| Using artificial intelligence search and pattern recognition techniques, new algorithms and methods were developed in two related areas of applied computer science: automated DNA restriction mapping, and automated DNA transcriptional enhancer element detection.; Two general techniques, using (i) the Dempster-Shafer theory of evidential reasoning, and (ii) a trained neural network, were developed for finding plausible solutions to systems of difference constraints, in cases where the differences are the observed values of random variables, instead of known constants.; A series of experimental restriction mapping algorithms were designed, implemented, and tested. Additional experimentation led to the development of an input preprocessing module which significantly speeds up searches, and an output post-processing module which enables users to analyze large solution sets and reduce their apparent complexity. These techniques were incorporated into a powerful public domain multiple-enzyme restriction mapping tool, Mapper, which will be distributed via the World Wide Web.; A series of experimental neural networks were created, to determine if neural networks could be trained to recognize transcriptional enhancer elements. Networks trained using {dollar}>{dollar}100 instances of a particular enhancer performed well, and a methodology was developed for producing enhancer-recognizing networks with an average false negative rate on the order of 0.1%, and an average false positive rate below 0.01%. |
