Program Development for Computer Simulation and Analysis for Multiplex Chromatography

In this report, the programs developed for the analysis of simulated data in multiplex chromatography are written in the ASYST language. The integration of the program modules into a menu-driven, user-friendly prototype program is outlined. The program is developed to allow for flexibility and easy upgrading of the code. This report examines the calculation procedures potentially useful in multiplex chromatography. Analysis of the mutliplex chromatograms is done by three mathematical methods, Cross-Correlation, Hadamard Transform and Fast Fourier Transform. The effects of experimental length/chromatogram length and decision sequences on correlation noise present in a single chromatogram, are discussed. As the number of data points increases, n, the improvement of the signal-to-noise ratio in the resulting single chromatograms is found to be a little less than the theoretical rate of (n)0.5. A non-random decision sequence is discovered to introduce drift and noise into the multiplex chromatogram. The influence of electronic white noise and electronic drift in the experimental data (multiplex chromatogram) on the resulting single chromatograms is explored. As electronic white noise increases, the signal-to-noise ratio in the resulting single chromatogram decreases. Electronic drift also introduces noise. Discussion of the best methods for given experimental conditions is based upon the signal-to-noise ratio in the resulting single chromatograms. Cross-Correlation is theoretically the fastest method but is greatly influenced by drift. Hadamard Transform is slower than Cross-Correlation but is not effected by drift. Fast Fourier Transform has the best signal-to noise ratio in the resulting single chromatogram but the peak magnitudes are not always reflective of the real heights. Fast Fourier Transform is slow but if written in machine language, is faster than Cross-Correlation or Hadamard Transform written in the ASYST language