2D NMR SPECTRA, IMAGING RECONSTRUCTION & PATTERN RECOGNITION

THE LONG TERM OBJECTIVES OF THIS PROJECT ARE: (1) TO DEVELOPAN ENTIRELY NEW COMPUTER SOFTWARE APPROACH FOR MAKING UNAMBIGUOUS SPECTRAL ASSIGNMENTS IN 2-DIMENSIONAL FT NMR OF SMALL TO MIDDLE-SIZED PROTEINS AND OTHER BIOMOLECULES, AND (2) TO USE THE SAME SOFTWARE APPROACH FOR ENHANCEMENT OF MR IMAGES AND FOR AUTOMATED FEATURE RECOGNITION. THESE TWO APPARENTLY DIFFERENT PROJECT AREAS ARE ACTUALLY CLOSELY RELATED TECHNICALLY, AS THE COMBINED MULTIVARIATE ANALYSIS/LOGIC PROGRAMMING APPROACH TO BE USED IS ESSENTIALLY IDENTICAL (WITH DIFFERENT RULES). THE TWO PROJECT GOALS EACH HAVE IMPORTANT BIOMEDICAL RAMIFICATIONS: (1) RECOGNITION AND ISOLATION OF EACH SPIN SYSTEM IN A COMPLEX 2D NMR SPECTRUM FROM A SMALL PROTEIN WILL GREATLY ENHANCE STRUCTURE STUDIES OF THESE MOLECULES; (2) APPLICATION TO MRI WILL IMPROVE QUALITY OF THESE EXPERIMENTS, OBVIATE SOME OF MANY EXPERIMENTAL DIFFICULTIES,AND SIGNIFICANTLY ENHANCE IMAGE INTERPRETATION WHERE AMBIGUITIES ARE PRESENT. THIS PROJECT HAS AN EXCELLENT CHANCE TO GIVE RISE TO A FAMILY OF SOFTWARE PRODUCTS OF MAJOR COMMERCIAL IMPORTANCE IN BOTH THE BIOMEDICAL AND CHEMICAL MARKETPLACES.