Recognition of environmental and genetic effects on barley phenolic fingerprints by neural networks

Through computational analysis of high-performance liquid chromatography (H PLC) traces we find correlations between secondary metabolites and growth c onditions of six varieties of barley. Using artificial neural networks, it was possible to classify chromatograms for which the varieties were fertili zed by nitrogen and treated by fungicide. For each variety of barley we cou ld also differentiate it from the others. Surprisingly, all these classific ation tasks could be solved successfully by a simple network with no hidden units. When adding to the methodology pruning of the network weights, we w ere able to reduce the set of peaks in the chromatograms and obtain a neces sary subset from which the growth conditions and differentiation may be dec ided. In some instances, more complex networks with hidden units could lead to a further reduction of the number of peaks used. In most cases, far mor e than half of the peaks are redundant. We find that it requires fewer info rmation-rich peaks to perform the variety differentiation tasks than to rec ognize any of the growth conditions. Analysis of the network weights reveal s correlations between weighted combinations of peaks. (C) 2001 Elsevier Sc ience Ltd. All rights reserved.