Analysis of protein and tannin contents in sorghum (Sorghum bicolor (L.) Moench) using near infrared reflectance spectroscopy
Shayo, Nicholas Benedict
The objective of this study was to determine whether near infrared reflectance spectroscopy (NIRS) could be effectively used to determine the protein and tannin content of sorghum. In addition, tannin content of whole and progressively dehulled sorghum grain was predicted using NIRS. The protein content of sorghum grain grown in 1983/84 and 1984/85 was analyzed by the Kjeldahl method and ranged from 7.4 to 16.7% and 8.4 to 15.1%, respectively. Tannin content was analyzed using five different methods; vanillin hydrochloric acid assay with an extraction time of 24 hr (V-HCl/24 hr), V-HCl/20 min, modified vanillin hydrochloric acid assay with an extraction time of 24 hr (MV-HCl/24 hr), MV-HCl/20 min, and Prussian blue assay with an extraction time of 10 min (PB/10 min). The correlation coefficients between the methods for tannin analysis ranged from 0.94 to 0.99. Protein content of sorghum grain was predicted with reasonable accuracy using NIRS. The standard error of prediction corrected for machine bias (SEPc) 0. 43%, ratio of range/SEPc (20. 8), and correlation coefficient (r) between the protein content analyzed by the Kjeldahl method and protein content predicted by NIRS was 0.99 (p < 0.01, n = 28) for the 1983/84 crop year; while the values for the 1984/85 crop year were SEPc (0.28%), ratio of range/SEPc (16.5) and r = 0.99 (p < 0.01, n = 20) . It was found that when the 1983/84 prediction equation was used to predict protein content in the 1984/85 crop year the SEPc was 0.35%, range/SEPc was 13.8, and r was 0.98 (p < 0.01, n = 20) while when the 1984/85 prediction equation was applied to the sorghum lines grown in 1983/84 the SEPc was 0.75%, range/SEPc was 12.4 and r was 0.99 (p < 0.01, n = 28). The SEPc and range/SEPc for the whole sorghum grain was 3.54% and 2.82, respectively. The correlation coefficient (0.66, p < 0.01, n = 20) between the V-HCl/24 hr assay and the NIRS method was low but significant. The technique appears not to be suitable for tannin analysis in whole sorghum grain. The correlation coefficient between tannin values obtained by five different assays and the NIRS method were 0.93, 0.91, 0.91, 0.86 and 0.86 for the V-HCl/24 hr, V-HCl/20 min, MV-HCl/24 hr, MV-HCl/20 min and PB/10 min, respectively; the SEPc were 1.44, 1.09, 0.67, 0.86 and 0.16% while the range/SEPc were 6.72, 6.74, 10.0, 9.84 and 7.75, respectively. The tannin content of sorghum lines with a narrow range of tannin (0 to 5%) was poorly predicted by NIRS with a correlation coefficient of 0.61 (p < 0.01, n = 20). The SEPc and range/SEPc was 1.32% and 3.55, respectively. Tannin content of two progressively dehulled sorghum lines was predicted with reasonable accuracy using NIRS. The SEPc was 0.11, range/SEPc was 19.1 and correlation coefficient between the tannin values obtained by the V-HCl/24 hr assay and those obtained by NIRS was r = 0.99 (p < 0.01, n = 20) for sorghum line IS9487 while sorghum line IS9613 gave a SEPc of 0 .35%, range/SEPc of 11.3 and a correlation coefficient of 0. 95 (p < 0.01, n = 26).