Identification of Geographical Origin of Chinese Chestnuts Using Hyperspectral Imaging with 1D-CNN Algorithm

The adulteration in Chinese chestnuts affects the quality, taste, and brand value. The objective of this study was to explore the feasibility of the hyperspectral imaging (HSI) technique to determine the geographical origin of Chinese chestnuts. An HSI system in spectral range of 400–1000 nm was applied to identify a total of 417 Chinese chestnuts from three different geographical origins. Principal component analysis (PCA) was preliminarily used to investigate the differences of average spectra of the samples from different geographical origins. A deep-learning-based model (1D-CNN, one-dimensional convolutional neural network) was developed first, and then the model based on full spectra and optimal wavelengths were established for various machine learning methods, including partial least squares-discriminant analysis (PLS-DA) and particle swarm optimization-support vector machine (PSO-SVM). The optimal results based on full spectra for 1D-CNN, PLS-DA, and PSO-SVM models were 97.12%, 97.12%, and 95.68%, respectively. Competitive adaptive reweighted sampling (CARS) and a successive projections algorithm (SPA) were individually utilized for wavelengths selection, and the results of simplified models generally improved. The contrasting results demonstrated that the prediction accuracies of SPA-PLS-DA and 1D-CNN both reached 97.12%, but 1D-CNN presented a higher Kappa coefficient value than SPA-PLS-DA. Meanwhile, the sensitivities and specificities of SPA-PLS-DA and 1D-CNN models were both above 90% for the samples from each geographical origin. These results indicated that both SPA-PLS-DA and 1D-CNN models combined with HSI have great potential for the geographical origin identification of Chinese chestnuts.

Application of Laser-Induced Breakdown Spectroscopy Coupled With Spectral Matrix and Convolutional Neural Network for Identifying Geographical Origins of Gentiana rigescens Franch

Accurate geographical origin identification is of great significance to ensure the quality of traditional Chinese medicine (TCM). Laser-induced breakdown spectroscopy (LIBS) was applied to achieve the fast geographical origin identification of wild Gentiana rigescens Franch (G. rigescens Franch). However, LIBS spectra with too many variables could increase the training time of models and reduce the discrimination accuracy. In order to solve the problems, we proposed two methods. One was reducing the number of variables through two consecutive variable selections. The other was transforming the spectrum into spectral matrix by spectrum segmentation and recombination. Combined with convolutional neural network (CNN), both methods could improve the accuracy of discrimination. For the underground parts of G. rigescens Franch, the optimal accuracy in the prediction set for the two methods was 92.19 and 94.01%, respectively. For the aerial parts, the two corresponding accuracies were the same with the value of 94.01%. Saliency map was used to explain the rationality of discriminant analysis by CNN combined with spectral matrix. The first method could provide some support for LIBS portable instrument development. The second method could offer some reference for the discriminant analysis of LIBS spectra with too many variables by the end-to-end learning of CNN. The present results demonstrated that LIBS combined with CNN was an effective tool to quickly identify the geographical origin of G. rigescens Franch.

Origin Identification of Hungarian Honey Using Melissopalynology, Physicochemical Analysis, and Near Infrared Spectroscopy

The objective of the study was to check the authenticity of Hungarian honey using physicochemical analysis, near infrared spectroscopy, and melissopalynology. In the study, 87 samples from different botanical origins such as acacia, bastard indigo, rape, sunflower, linden, honeydew, milkweed, and sweet chestnut were collected. The samples were analyzed by physicochemical methods (pH, electrical conductivity, and moisture), melissopalynology (300 pollen grains counted), and near infrared spectroscopy (NIRS:740–1700 nm). During the evaluation of the data PCA-LDA models were built for the classification of different botanical and geographical origins, using the methods separately, and in combination (low-level data fusion). PC number optimization and external validation were applied for all the models. Botanical origin classification models were >90% and >55% accurate in the case of the pollen and NIR methods. Improved results were obtained with the combination of the physicochemical, melissopalynology, and NIRS techniques, which provided >99% and >81% accuracy for botanical and geographical origin classification models, respectively. The combination of these methods could be a promising tool for origin identification of honey.

Chromatin Accessibility Profiling to Increase Diagnostic Accuracy and Refine Cell-of-Origin Classification of Mature T-Cell Lymphomas

Background: Mature T-cell lymphomas and leukemias (MTCL) are heterogeneous diseases with dismal prognosis. Differentiating between the numerous entities requires specialized pathology expertise and studies show up to 20% change in diagnosis after expert review of cases (Laurent, JCO, 2017). Assay for transposase accessible chromatin sequencing (ATAC-seq) is a simple technique to profile open chromatin regions (OCR) proven to be highly discriminant for cell-of-origin identification regardless of cell activation status (Shih, Cell, 2016). We applied ATAC-seq to MTCL in order to explore the epigenetic landscape of these diverse entities, compared them to normal T-cell subtypes and built a predictive model to help diagnosis. Method: Ten-thousand FACS-sorted single cells from primary MTCL samples and 50µm section of frozen tumoral tissue from the TENOMIC French T-cell Lymphoma Consortium were processed according to the previously published FAST-ATAC and OMNI-ATAC protocols respectively (Corces, Nat Genetics, 2016 & Nat. Methods, 2017). Concurrently we applied FAST ATAC to different normal T- and NK-cell subsets sorted from healthy donor PBMC or lymph node suspensions. Sequencing data were processed by an adapted version of ENCODE ATAC-seq pipeline. Matrix of insertion events in peaks by sample was obtained, normalized and most variant peaks were selected for UMAP projection. Results: In total, 678 normal and tumoral samples were sequenced to provide a comprehensive landscape of chromatin accessibility in MTCL. Epigenetic profiling by ATAC-seq of FACS-sorted tumoral samples resulted in a complete segregation of the known MTCL entities (AITL, TFH-PTCL, ALK+ and ALK- ALCL, HSTL, CTCL, ATLL, LGL and T-PLL). Most PTCL-NOS (13/17) clustered with a pre-defined MTCL subtype (mainly AITL/TFH-phenotype PTCL, CTCL and lymphomas exhibiting cytotoxic features). All but one discordant diagnosis between pathology and ATAC-seq (1/11) led to revised diagnosis after pathology review. Unsupervised clustering of normal NK- and T-cell subtypes (N=49) and sorted tumoral lymphoma cells (N=104) confirmed that AITL derive from TFH cells. HSTL and LGL closely segregated with NK- and gamma-delta T cells, in line with their known innate-like phenotype. Surprisingly, the cell-of-origin of T-PLL seems to be naïve T cells despite the known expression of central memory markers on leukemic cells. Beyond epigenetic classification, background reads from ATAC-seq profiles were used to detect copy number variation (CNV), such as isochromosome 7q in HSTL. In addition, HTLV1 and EBV viral sequence detection in ATAC-seq reads strengthened identification of ATLL and NKTCL cases. Finally, using unsupervised deconvolution approaches, we were able to discriminate different MTCL subtypes from 223 processed bulk frozen samples. All known MTCL subtypes were differentiated (AITL/PTCL-TFH, HSTL, NKTCL, ATLL, ALK- and ALK+ ALCL, MEITL, EATL). A subgroup of PTCL-NOS harboring GATA3 OCRs and a distinctly high CNV number was isolated that might correspond to previously described PTCL-GATA3 subtypes (Iqbal, Blood, 2019). A random forest model was trained to predict diagnosis based on chromatin-accessibility clusters defined in the discovery cohort of patients. The model showed accurate prediction performance by cross-validation. External validation on 172 samples collected from 5 tertiary care centers will be presented at the meeting. Conclusion: ATAC-seq is a fast and cost-effective technique to help and refine MTCL pathological classification and allows for putative cell-of-origin identification in lymphoma. Training of a machine learning model to predict MTCL entity diagnosis based on ATAC-seq analysis of fresh or frozen samples shows promising results. Figure 1 Figure 1. Disclosures Sibon: Janssen: Consultancy; Abbvie: Consultancy; iQone: Consultancy; Takeda: Consultancy; Roche: Consultancy. Drieux: Genexpath: Patents & Royalties: The author is a potential inventor on a patent application for the LymphoSign, which has been licensed for by Genexpath Patents & Royalties.. Ruminy: Genexpath: Patents & Royalties: The author is a potential inventor on a patent application for the LymphoSign, which has been licensed for by Genexpath Patents & Royalties. . Salles: Takeda: Consultancy; Velosbio: Consultancy; Ipsen: Consultancy; Allogene: Consultancy; Miltneiy: Consultancy; Genentech/Roche: Consultancy; Genmab: Consultancy; Janssen: Consultancy; Loxo: Consultancy; Kite/Gilead: Consultancy; Regeneron: Consultancy, Honoraria; Morphosys: Consultancy, Honoraria; Novartis: Consultancy; Incyte: Consultancy; Rapt: Consultancy; Epizyme: Consultancy, Honoraria; Debiopharm: Consultancy; BMS/Celgene: Consultancy; Beigene: Consultancy; Abbvie: Consultancy, Honoraria; Bayer: Honoraria. Gaulard: Alderaan: Research Funding; Sanofi: Research Funding; Innate Pharma: Research Funding; Gilead: Consultancy; Takeda: Consultancy, Honoraria. Bachy: Kite, a Gilead Company: Honoraria; Novartis: Honoraria; Daiishi: Research Funding; Roche: Consultancy; Takeda: Consultancy; Incyte: Consultancy.

Application of DNA Barcoding for the Identification of a Traditional Chinese Medicine Shedan

Background Shedan has a long history of application in Traditional Chinese medicine (TCM), however, Shedan from different original source has been indiscriminately used. So far, there is still a lack of an effective tool to differentiate the original source of Shedan medicinal materials, which brings great risk to the safety and effectiveness of clinical applications. Hence, it is imperative to develop a practicable approach to identify Shedan medicinal materials. Methods The specificity of two pairs of primers, including Folmer’s universal primers and a pair of originally designed primers COISNFF/COISNFR, was tested to screen the more specific primers for further origin identification of Shedan. A total of 253 fresh snake gallbladder samples from 31 morphologically identified snake species were collected and authenticated. Moreover, 51 fresh snake bile samples and 17 fresh bile samples from five other common domestic poultry and livestock (cattle, chicken, duck, pig and sheep) were collected and distinguished using the more specific primers. Additionally, a total of 195 market Shedan samples randomly selected from 18 batches of Shedan medicinal materials were investigated. Sequence definition was executed by querying sequence similarities in GenBank and the Barcode of Life Data System (BOLD), respectively. Results It turned out that the standard COI barcode obtained by COISNFF/COISNFR primers, rather than Folmer’s universal primers, can distinguish all the testing samples from each other in fresh Shedan samples, and COISNFF/COISNFR primers were also specific to snake species and the other four animal species except duck. In terms of market Shedan, 84.6% (165/195) samples can be attributed to 13 snake species from four families and 4.6% (9/195) can be attributed to adulterated chicken species. Conclusion The COI-based DNA barcoding was practicable for species identification of Shedan used in traditional Chinese medicine. The original source of current market Shedan, including adulterated species, has been preliminarily clarified, which provides a foundation for quality control of Shedan medicinal materials.

Authentication of Geographical Origin in Hainan Partridge Tea (Mallotus obongifolius) by Stable Isotope and Targeted Metabolomics Combined with Chemometrics

Partridge tea (Mallotus oblongifolius (Miq.) Müll.Arg.) is a local characteristic tea in Hainan, the southernmost province of China, and the quality of partridge tea may be affected by the producing areas. In this study, stable isotope and targeted metabolomics combined chemometrics were used as potential tools for analyzing and identifying partridge tea from different origins. Elemental analysis—stable isotope ratio mass spectrometer and liquid chromatography-tandem mass spectrometrywas used to analyze the characteristics of C/N/O/H stable isotopes and 54 chemical components, including polyphenols and alkaloids in partridge tea samples from four regions in Hainan (Wanning, Wenchang, Sanya and Baoting). The results showed that there were significant differences in the stable isotope ratios and polyphenol and alkaloid contents of partridge tea from different origins, and both could accurately classify partridge tea from different origins. The correct separation and clustering of the samples were observed by principal component analysis and the cross-validated Q2 values by orthogonal partial least squares discriminant analysis (OPLS-DA) were 0.949 (based on stable isotope) and 0.974 (based on polyphenol and alkaloid), respectively. Potential significance indicators for origin identification were screened out by OPLS-DA and random forest algorithm, including three stable isotopes (δ13C, δ D, and δ18O) and four polyphenols (luteolin, protocatechuic acid, astragalin, and naringenin). This study can provide a preliminary guide for the origin identification of Hainan partridge tea.

Quality Assessment and Origin Identification on Four Species of Curcumae Radix by HPLC Assay and Fingerprint Analysis

BackgroundCurcumae Radix is a multi-source Chinese medicine called Yujin. It is derived from the root tubers of four Curcuma species including Curcuma wenyujin , C. kwangsiensis , C. phaeocaulis and C. longa. The identification of the root tubers of the four Curcuma species often caused confusions. MethodsIn this study, we developed HPLC/PDA methods to differentiate these four species of Curcumae Radix by HPLC assay and fingerprint analysis. The methods developed were also validated with their precision, repeatability and accuracy. ResultsCurdione could only be detected in C. wenyujin . For C. kwangsiensis , C. phaeocaulis , if the relative peak area value of peak 1 and germacrone peak of the samples was more than 1.024, it could be identified as C. phaeocaulis of Curcumae Radix. Curcumin, desmethoxycurcumin and bisdesmethoxycurcumin could only be detected in C. longa . We also developed a simple method to determine curcumin, desmethoxycurcumin and bisdesmethoxycurcumin in C. lon ga and germacrone in the other three species of Curcumae Radix. ConclusionsBecause of the definite difference of main ingredients between curcumin-free group (C. wenyujin , C. kwangsiensis and C. phaeocaulis ) and curcumin containing C. longa from this study, four Curcuma species of Curcumae Radix in the pharmacopeia should be separated into two monographs, with curcumin-free and germacrone containing species as one and curcumin containing C. longa as the second monograph. The HPLC/PDA methods of HPLC assay and fingerprint analysis developed in this study could be used for the quality control of these four species of Curcumae Radix. Keywords: Curcumae Radix, HPLC, assay, fingerprint, curcumin, desmethoxycurcumin, bisdesmethoxycurcumin, germacrone, curdione