Near infrared reflectance spectroscopy for protein content in soybean flour and screening of germplasm across different countries

Abstract

Maintaining high protein content in soybean seeds is critical in view of the premium offered for protein content in soymeal in international market. Besides, quality of several soy-based products depends upon crude protein content of the initial raw material used. Determination of protein content in soybean seeds through wet chemistry involves use of chemicals and is time-consuming and labour-intensive. In the present investigation, a calibration model for determination of protein content in ground soyflour using non-destructive method of near-infrared reflectance spectroscopy (NIRs) was developed. A set of samples with wide variation in protein content as measured through wet chemistry were scanned by NIRs between 1100 and 2100 nm at 2-nm interval. High determination coefficient (R2) and low values for root-meansquare error as well as standard error of prediction of cross-validation confirmed the utility of model for prediction of future samples. For external validation, a set of genotypes from different genetic background were tested and were predicted with good accuracy. Lower values of standard error of calibration and prediction were observed than reported previously for whole seeds. Subsequently, 1210 soybean accessions from 18 countries were screened through NIRS. Average protein content of soybean accessions from different countries varied significantly. Genotypes identified for high protein content from different countries may be used for development of genotypes containing further high protein content.