Effect of rice variety and blending proportion on the proximate compositions, minerals and phytic acid contents of bread from rice-teff blend

Development of bakery products containing rice ( Oryza sativa, Linn.) and teff ( Eragrostis tef) could have potential health benefits due to their gluten free nature. Nine experimental runs were generated using custom design by JMP 8 software. The effect of two factors, rice variety (Edeget, X-jigna and Nerica-4) and blending proportions of rice and teff (0.5:0.5, 0.7:0.3 and 0.9:0.1) were studied. The data analysis was conducted using SAS software package for the mean comparison and custom design by JMP 8 software. Response surface methodology was applied to study the interaction effect of the main factors and to generate the predictive equations. An optimal value (1.60%) of fiber was obtained when the proportion of the blend was 50% Edeget and 50% teff because teff grain is high in fiber. A maximum value (10.75%) of protein was obtained when the proportion of the blend was 70% Nerica-4 and 30% teff. Carbohydrate was optimal (81.37%) when 90% Edeget and 10% teff were blended because rice grain is high in carbohydrate. Optimal iron content (12.97 mg/100g) was obtained when the proportion of the blend was 50% Nerica-4 and 50% teff because teff grain is high in iron. Optimal zinc content (4.14 mg/100g) was obtained when the proportion of the blend was 50% X-jigna and 50% teff. The optimal value (61.25 mg/100g) of calcium was obtained when the proportion of the blend was 50% Edeget and 50% teff. Optimum (lower) value (0.31mg/g) of phytic acid was obtained when the proportion of the blend was 90% Nerica-4 and 10% teff because rice grain is lower in phytic acid content. It was concluded that rice variety and rice-teff blending proportion had a significant effect on the physico-chemical properties of rice-teff blend bread. An optimal nutrient blend (high in nutrients, low in anti-nutrients) was obtained when 70% Edeget rice variety was blended with 30% teff. All the derived mathematical models for the various responses were found to fit significantly to the predicted data.