High throughput phenotyping of Camelina sativa seeds for crude protein, total oil, and fatty acids profile by near infrared spectroscopy

Fast, non-destructive methods for determining the seed composition of Camelina sativa (L.) Crantz would be beneficial in evaluating germplasm for important agronomic traits. In this study, near infrared spectroscopy (NIRS) methods were developed and evaluated for conducting non-destructive, high throughput phenotyping of seed quality traits. Crude protein and total oil content for 85 accessions (63 summer- and 22 winter-biotypes) were first determined by established wet chemistry methodology; whereas, for fatty acid profiles 173 accessions (149 summer- and 24 winter-biotypes) were determined using Gas Chromatography (GC). The wet chemistry and GC data were used to develop NIRS calibration equations for each trait. Based on the wet chemistry data obtained from 85 accessions, mean crude protein content was significantly less in summer (300 g kg−1) than in winter (315 g kg−1) biotypes (P ≤ 0.05) and total oil was greater in seeds of summer (351 g kg−1) than that of winter (326 g kg−1) biotypes. Coefficient of determination (r2 = 0.979 and 0.894, respectively) and ratio of performance to deviation (RPD = 9.15 and 4.33, respectively) for crude protein and oil content indicated a high level of confidence for predicting these traits using NIRS. Evaluation of all 173 accessions by NIRS did not appreciably change the predicted mean crude protein content of summer- and winter-biotypes; however, it did change the predicted mean total oil content of summer biotypes (260 g kg−1), which was significantly less than predicted for winter biotypes (323 g kg−1). Fatty acids contents were not significantly different between summer- and winter-biotypes. The most abundant fatty acid was linolenic acid (18:3) ranging from 22.8 to 38.4%, followed by linoleic acid (18:2) at 15.2–27.1%, eicosenoic acid (20:1) at 11.6–18.2%, and oleic acid (18:1) at 9.1–22.1%. Calibration models for the main fatty acids oleic, linoleic, linolenic, and eicosenoic acids had r2 values of 0.718, 0.790, 0.828, and 0.586, respectively. Results of this study indicate that NIRS has potential as a non-destructive, high throughput method for determining quality traits of camelina seed.