Geographical origin modeling of Chinese rice using stable isotopes and trace elements

Rice is an important staple food in China, which authenticity is closely associated with nutrition and safety. It is necessary to discriminate their geographical origin with a comprehensive databank. 900 Japonica and Indica rice samples from 17 provinces were collected to analysis stable isotopes and trace elements for their origin discrimination of four regions as Middle-Lower Yangtze Plain (Y-R), northeast (N-E), southwest (S–W) and southeast (S-E). Results revealed Japonica rice was isotopically more positive than Indica rice, becoming −27.3‰ vs. −28.5‰ for δ 13 C, 5.2‰ vs. 4.6‰ for δ 15 N, −58.7‰ vs. −65.1‰ for δ 2 H and 20.3‰ vs. 18.1‰ for δ 18 O in Japonica vs. Indica rice, respectively. N-E rice had the most positive stable isotope values and Y-R rice had higher elemental contents. Using neural network modeling, Japonica rice from N-E and Y-R was discriminated with high accuracy of 97.2%. And Indica rice could also be geographically assigned to Y-R, S–W and S-E with the accuracy of 76.0% for blind samples. This study establishes the first comprehensive stable isotopic and elemental geographical database for Chinese rice and provides a promising discrimination method to key production regions in China. • Isotopes and elements in rice are profiled across key producing regions in China. • Geographical and varietal characteristics of Japonica and Indica rice are studied. • Origin of Japonica and Indica rice were classified using neural network models. • Rice from northeast region could be differentiated with high accuracy >97%.