High-throughput automated post-processing of separation data.

Abstract

The development of an efficient method for high-throughput analysis of multiple electropherograms or chromatograms collected in series is presented. The method, encoded in a computer program designated "Cutter", utilizes batch processing for determining chromatographic figures of merit (CFOM) including peak centroid times, heights, areas, signal-to-noise ratios (S/N), variance (sigma2), skew, excess, and plate number (N) across a set of separations collected serially. The software was validated using simulated data with varying S/N, skew, and excess. The accuracy of the analysis was comparable to or improved over commercial software with area calculation relative errors (RE) below 5% for simulated data with S/N = 5. File sets containing 1300 electropherograms were analyzed in 5 min, representing a nearly 200-fold reduction in analysis time from other methods. Incorporated within the program is a novel method for automated peak deconvolution using an Empirically Transformed Gaussian function. Area measurements of deconvoluted peaks were within 3% of the true value of a simulated data set with S/N = 5 and resolution (R(S)) = 1 for equivalent peaks, and within 10% when the ratio of the overlapped peak heights was 10:1.