Rapid Prediction of Camelina Seed Oil Content Using Near-Infrared Spectroscopy

Camelina is a promising feedstock as a result of its ability to provide high-quality edible oil and jet fuel. However, predicting its oil content currently requires time- and labor-intensive analyses. Near-infrared (NIR) spectroscopy provides a rapid, low-cost determination approach for oil seed characterization. The objective of this study was to develop a NIR model to predict the camelina oil content using 200 camelina seed simples. Partial least squares (PLS) regression and principal component regression (PCR) were used to compare the performance of calibration models to the full spectra range (4000–1000 cm–1). PLS regression showed better prediction performance than PCR. The optimal model provided excellent fitness, with a R2 of 0.94 and root mean square of prediction error of 0.495%, making the model useful in various applications, including quality assurance and screening. This study confirmed that the NIR method significantly reduces time (from 60 to 1 min) and cost [from 20 to 1 United States dollar (USD)] required to determine the camelina seed oil content. Last but not least, this study leverages a high-throughput, cost-effective prediction method of the camelina oil content to facilitate plant breeding and genetic studies. Future work on the NIR model should focus on developing a model system to achieve rapid analysis of genomics at a low cost to assist plant feedstock improvement.