| 摘要 |
PROBLEM TO BE SOLVED: To improve reliability in the case of searching for a marker for distinguishing between two sample groups by utilizing discriminant analysis.SOLUTION: A peak matrix is created from mass spectral data on respective samples belonging to a plurality of sample groups (S1, S2), and PLS-DA is executed to the peak matrix to calculate loading values for every peak S3. In addition, an average value and standard deviation of peak values of the respective peaks are calculated to calculate CV values for every sample group from the peak matrix S4, and a loading correction value is calculated by dividing the loading values by product of the CV values for every sample group at the peaks for every peak S5. Since the CV values become smaller when variation of the peak values at the same m/z value is large and when the peak values are excessively large in only a specific sample, since the correction values become smaller even if the loading values are large, and search for a marker with high reliability becomes possible by being based on the loading correction values of the respective peaks.SELECTED DRAWING: Figure 2 |
