DNA sequence determination has become a standard method for the characterization of newly isolated genes. Recent advances in the technology for the determination of DNA sequences have allowed large amounts of data to accumulate rapidly. However, methods for transferring that data into the computer, where it is to be analysed, are still far from ideal, in that they require considerable manual interention. We propose to develop a highly automated system that will greatly increase the accuracy and the speed with which sequence data can be collected and analysed. The method requires first producing a digital image of the sequencing radioautogram. Algorithms will then be devised that are able to convert that ditigized data into a DNA sequence. A key feature of the software for this interpretative step is that it will operate according to a set of heuristic procedures that are clearly defined and which can easily be modified and extended. When these rules fail to provide an unambiguous interpretion then operator assistance will be requested to make a final decision. An important consequence of the strategy is that the primary data upon which the DNA sequence depends will be available in machine-readable form. Appropriate software will be developed to manage this primary data and to assist in the resolution of discrepancies that may arise during the course of a sequencing project. The software tools that will be developed during the course of this project would have additional applications for the interpretation of all kinds of biological data that derive from gel electrophoretic techniques.
