Additional file 1 of Machine learning integration of multi-modal analytical data for distinguishing abnormal botanical drugs and its application in Guhong injection

Additional file 1: Figure S1. Chromatographic fingerprints of 78 batches of GHIs based on HPLC-UV. Figure S2. The permutation test to validate the validity of PLS-DA based on RPA of HPLC-UV fingerprints. Figure S3. Chromatographic fingerprints of 78 batches of GHIs based on HPLC-ELSD. Figure S4. The permutation test to validate the validity of PLS-DA based on RLPA of HPLC-ELSD fingerprints. Figure S5. q1HNMR spectra of 78 batches of GHIs. Figure S6. 13C NMR spectrum of GHI in CD3OD. Figure S7. HSQC spectrum of GHI in CD3OD. Figure S8. COSY spectrum of GHI in CD3OD. Figure S9. HMBC spectrum of GHI in CD3OD. Figure. S10. Comparisons between the spectra of reference compounds and Guhong Injection. Figure. S11. Deconvolution resolved signal overlap for q1HNMR quantitation of Valine. Figure S12. The permutation test to validate the validity of PLS-DA based on the content of compounds by q1HNMR. Table S1. Label, Lot number and Volume of 52 normal batches of GHIs. Table S2. Label, Lot number and Volume of 26 abnormal batches of GHIs. Table S3. RPA extracted from HPLC-UV fingerprints of 78 batches of GHIs. Table S4. Pearson correlation coefficient based on RPA between 78 batches of GHIs. Table S5. RLPA extracted from HPLC-ELSD fingerprints of 78 batches of GHIs. Table S6. Pearson correlation coefficient based on RLPA between 78 batches of GHIs. Table S7. Absolute concentrations of compounds from 78 batches of GHIs (mg·mL-1). Table S8. Pearson correlation coefficient based on content between 78 batches of GHIs. Table S9. Fused feature table. Table S10. Trainset based on fused feature table. Table S11. Testset based on fused feature table. Method validation S1. Method validation of HPLC-UV fingerprints. Method validation S2. Method validation of HPLC-ELSD fingerprints.