Access Keys:
Skip to content (Access Key - 0)

General NIRS

Skip to end of metadata
Go to start of metadata
  • None

Near-Infrared Reflectance Spectroscopy (NIRS) Analysis

NIRS is a rapid and relatively inexpensive technique that facilitates the analysis of several traits simultaneously. The advantages for using NIRS in screening and breeding programs are various:
  • Routine application of NIRS does not need chemical reagents and avoids contamination with chemical waste.
  • Fast analysis of several traits simultaneously in less than two minutes per sample are possible and several hundred samples can be analyzed per day.
  • The method needs only simple sample preparation like drying and milling in sweetpotato, potato and cassava. In general the method is non-destructive, so that intact grains of rice, wheat, pearl millet, beans and maize can be analyzed by NIRS and later can be sown.
  • In application it is a cheap method analyzing several traits simultaneously.
The NIRS technique is a spectroscopy method using the wavelength range of 400-2500nm. The monochromatic light is diffusely reflected to the detector, as a result of reflection, refraction and diffraction on particles and droplet surfaces inside the sample. The spectrum is a result of specific light absorptions for example from C-H, O-H, N-H, C=O and combination bands at different wavelengths. NIRS is an indirect method. Owing to non-selectivity in NIR spectra, a multivariate calibration of a number of spectra against values from reference chemical methods is needed. A critical issue for reliable, robust and precise calibrations is the selection of "calibration samples". It is important that the calibration samples span the variation (different cultivars/genotypes, regional differences, environments etc.) of the target population sample to predict. Each calibration sample is subjected to both spectroscopic and chemical reference method measurements before combining them. To start using NIRS you have to develop calibrations with reference laboratory values of a minimum of 100-120 samples. The estimations must be robust, so the calibration dataset should include a wide range of genotypes grown in different environments. Calibration equations must be validated and extended frequently with new samples.

NIRS across major food crops in the HarvestPlus