Discrimination of transgenic cotton seed using visible and near-infrared diffuse reflectance spectroscopy (NIRS)

Visible/near-infrared (Vis/NIR) diffuse reflectance spectroscopy combined with chemometrics techniques, was used to distinguish transgenic cotton seed from non-transgenics. Two hundred fifty cotton seeds of RCH-2 genotype containing Cry 1Ac gene conferring resistance to lepidopteron pests and the same number of their parent non-transgenic seeds were scanned in the Vis/NIR wavelength spectrum of 400–2500 nm. Modified partial least square (mPLS), partial least square (PLS) and principal component regression (PCR) models were applied for calibration and classification of samples into two groups. The results showed that differences exist between transgenic and non-transgenic cotton seeds and excellent classification can be obtained after optimizing spectral pretreatment. The spectral difference between the two groups are observed at a wavelength range of 1100–1900 nm, which is related to first and second overtone of C-H stretching vibrations and sixth overtone of C=C stretching vibrations. Standard normal variate (SNV) and detrend scatter correction with second derivative data pretreatment using mPLS model could achieve 100% accurate classification for both transgenic and non-transgenic samples. Reliable equations were developed with r = 0.96 and r = 0.92 for calibration and validation set respectively with low standard error of performance (SEP) (0.13) using mPLS model.