A perceptron type neural network simulator for structure-activity correlation of molecules, Neco, has been developed with two different pre-education methods : back propagation method and reconstruction of weight matrix method. Since the program was written in C, it is executable on popular Unix workstations. The number of intermediate layered units in intermediate layer and intermediate layer itself can be set arbitrarily by the user. As an example of application, conformations of norbornene isomers were predicted using 13C-NMR data. The predicted conformations are in excellent agreement with experiments. By reconstructing the weight matrix, it was suggested that 13C -NMR data of only two specified carbons in the norbornene skeleton have strong correlation with the conformation (exo and endo) of the main branch. This was verified by the parameter scan method.
A new quantity, stickiness, was proposed that enables us to deal quantitatively with interactions between a macromolecule (e.g., DNA) and a ligand (e.g., oligonucleotide) and gives an image of global/local effects in molecular interactions.Pseudostickiness, its manageable version based on binding constant, was calculated and applied to an analysis of lambda DNA sequence, revealing some characteristics of the lambda genome. The utility of stickiness in understanding specific interactions of biomolecules is also discussed.
Microcomputer-aided learning (CAL) has been put into effect in various ways in chemistry teaching. Computer support is a useful tool for schooling. Students learn in detail about the experimental contents by themselves before beginning an experiment in the desired study, and then may carry out microcomputer-aided learning. We enforce learning before the experimental training in our College, and examine the training effect on education. A teaching materials program for learning was developed by graduation study work. There are 18 themes (physical chemistry experiments). Students must understand the experimental operation necessary for beforehand training in every theme. We have been developing a beforehand learning program (CAL software) in each theme (physical chemistry experiment). A student beginning an experiment uses the program and then puts the training into effect. We have examined the CAL training effect on education for cases with or without beforehand learning and found that such learning has a marked positive impact on the experimental training of the student in the chemistry curriculum.