A Neural Network That Classifies Mass Spectra

mass spectral classification; structure elucidation; neural networks; back propagation We have designed a feed-forward neural network to classify low-resolution mass spectra of unknown compounds according to the presence or absence of 100 organic substructures. The neural network, MSnet, was trained to compute a maximum-likelihood estimate of the probability that each substructure is present. We discuss some design considerations and statistical properties of neural network classifiers, and the effect of various training regimes on generalization behavior. The MSnet classifies mass spectra more reliably than other methods reported in the literature, and has other desirable properties.

[1]  M. Koehler,et al.  Application of pattern recognition to mass spectral data of toxic organic compounds in ambient air , 1987 .

[2]  Yann LeCun,et al.  Improving the convergence of back-propagation learning with second-order methods , 1989 .

[3]  Stephen R. Lowry,et al.  Comparison of various K-nearest neighbor voting schemes with the self-training interpretive and retrieval system for identifying molecular substructures from mass spectral data , 1977 .

[4]  F. Bordwell.  Determination of organic structures by physical methods (Nachod, F. C.; Phillips, W. D.; eds.) , 1962 .

[5]  E. Feigenbaum,et al.  Applications of artificial intelligence for chemical inference—X , 1973 .

[6]  David E. Rumelhart,et al.  Predicting the Future: a Connectionist Approach , 1990, Int. J. Neural Syst..

[7]  P. T. Palmer,et al.  Development of algorithms for automated elucidation of spectral feature/substructure relationships in tandem mass spectrometry , 1988 .

[8]  J. Ross Quinlan,et al.  Learning Efficient Classification Procedures and Their Application to Chess End Games , 1983 .

[9]  S. Wold,et al.  Extraction of mass spectral information by a combination of autocorrelation and principal components models , 1984 .

[10]  E. Feigenbaum,et al.  Applications of artificial intelligence for chemical inference. III. Aliphatic ethers diagnosed by their low-resolution mass spectra and nuclear magnetic resonance data , 1969 .

[11]  Peter G. Anderson,et al.  The interpretation of i.r. and Raman spectra using pattern recognition , 1977 .

[12]  Geoffrey E. Hinton,et al.  Learning representations by back-propagating errors , 1986, Nature.

[13]  Thomas L. Isenhour,et al.  Information content of mass spectra as determined by pattern recognition methods , 1974 .

[14]  Hervé Bourlard,et al.  Generalization and Parameter Estimation in Feedforward Netws: Some Experiments , 1989, NIPS.

[15]  N. A. B. Gray,et al.  Constraints on "learning machine" classification methods , 1976 .

[16]  K. Varmuza.  Pattern recognition in analytical chemistry , 1980 .

[17]  F. McLafferty,et al.  Computer‐aided interpretation of mass spectra , 1969 .

[18]  K. Varmuza,et al.  Selective detection of classes of chemical compounds by gas chromatography/mass spectrometry/pattern recognition: polycyclic aromatic hydrocarbons and alkanes , 1987 .

[19]  Lorien Y. Pratt,et al.  Comparing Biases for Minimal Network Construction with Back-Propagation , 1988, NIPS.

[20]  Terrence J. Sejnowski,et al.  Learned classification of sonar targets using a massively parallel network , 1988, IEEE Trans. Acoust. Speech Signal Process..

[21]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[22]  John Moody,et al.  Fast Learning in Networks of Locally-Tuned Processing Units , 1989, Neural Computation.

[23]  P. Harrington,et al.  Approaches to Pyrolysis/Mass Spectrometry Data Analysis of Biological Materials , 1990 .