Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The effect of genotype and traditional food processing methods on in-vitro protein digestibility and micronutrient profile of sorghum cooked products


Sorghum (Sorghum bicolor (L.) Moench) is one of the principal staple for millions of people in sub-Saharan Africa serving as the main sources of protein. However, protein digestibility is low in sorghum and this may be affected by processing methods. In this study 15 sorghum cultivars and one variety each of maize (Zea maize) and tef (Eragrostis tef) all of Ethiopian origin were investigated for in-vitro protein digestibility (IVPD), activity and concentration of anti-nutritional factors and micro nutrient profile in raw flour and various cooked food samples. Kafirin composition content and composition was also determined from raw flour samples of the sorghum cultivars. IVPD was significantly different between genotypes with both maize and tef superior to sorghum both in cooked and uncooked state except for the high lysine genotype Wetet Be-gunchie. Cooking significantly reduced IVPD in all crops but had only minor effect in maize. Results revealed a highly significant interaction between genotype and food processing methods where, occasionally, genotypes with highest IVPD under one processing method ended up to be the lowest under another. Trypsin inhibitor levels had a significant and negative correlation with IVPD (r2 = 0.1), while changes in phytic acid concentration and intrinsic phytase levels during processing followed opposite trends to each other. Processing increased mineral levels by 20–44% for iron and 4–29% for zinc perhaps due to degradation of phytic acid. Results demonstrated that protein digestibility and the concentration of anti- nutritional factors varied widely depending on the food type. Identification of specific genotypes for a specific food product may help improve the nutritional quality of sorghum based foods.


An important food crop in the semi-arid tropics of Asia and Africa, sorghum [Sorghum bicolor (L.) Moench] is also a key feed grain in many other parts of the world. Though similar to other cereals in its grain composition [1], the nutritional quality of sorghum is impacted by several endogenous and exogenous factors that make its proteins less digestible than other cereals, especially when wet-cooked [27]. Past research efforts made to address this problem have focused mainly on understanding the role of protein body structure [8,9,10] and its chemistry [4,8] producing valuable information that served as foundation for numerous research initiatives.

Research to determine the effect of various food-processing methods and anti-nutritional compounds on sorghum protein digestibility has also been carried out. However, the impacts have been largely genotype dependent and studies based on limited number of genotypes may not provide the whole picture of the complexities surrounding sorghum protein digestibility. Anti-nutritional compounds identified to be of significant interest for protein digestibility include phytic acid (Myo-inositol-1,2,3,4,5,6-hexakisphosphate), trypsin protease inhibitors, and phenolic compounds such as tannins. These compounds are reported to interfere with protein digestibility in one way or another and limit its protein bioavailability [6,11]. Phytic acid (phytate) acts as the major storage form of phosphorous accounting for 1–5% by weight in legumes, cereals, oil seeds, and nuts [12,13]. While phytate-protein complexes make proteins less digestible, phytate also chelates iron and zinc hindering both macro and micronutrient bioavailability [14,15,16]. Hence, low phytate foods or food products with degraded phytate have been shown to improve iron absorption [16,17] and should also increase absorption of other micronutrients.

Protease inhibitors, are another important group of anti-nutritional factors undermining protein digestion in grains and food products. Due to their abundance in leguminous seeds, trypsin inhibitors have been extensively studied in legumes including soybean [1821] and lima bean [22,23,24]. Protease inhibitors are also reported in several cereals including maize [25,26] wheat [27], barley [27,28, 29], rye [27], oats [30], rice [31,32] and sorghum [33,34]. Sorghum is known to possess three iso-forms of trypsin protease inhibitors labelled as inhibitor I through III based on their biochemical properties. Inhibitors I and II inhibit both trypsin and chymotrypsin; whereas inhibitor III is known to primarily inhibit chymotrypsin [34].

Food processing methods influence the bioavailability of nutrients in cereal grains either through altering the effects of anti-nutritional factors or due to chemical changes that may occur to the proteins and its complexes with other compounds [3542]. The best-known example of this for sorghum is the decrease in protein digestibility when sorghum is wet-cooked [9]. Other forms of cooking such as dry heating (e.g. popping) also have a negative impact on protein digestibility but to a lesser degree [43,44]. Food processing techniques such as extrusion, fermentation, dry roasting, malting/germination also affects digestibility of sorghum proteins [45,46]. However, many of previous studies have used laboratory based cooking methods and investigated only limited number of genotypes and thus may not adequately the actual food products. Scaling up the research to include larger number of genotypes and food products typically produced and consumed by native consumers of the crop may provide not only a robust data but also array of information that may be of use for disentangling the complexity related to sorghum protein digestibility. Such information may also be of critical importance in guiding the choice of germplasm for breeding new cultivars suited for making specific food products or those for broader food applications. Therefore, the present study was aimed at evaluating the nutritional value of diverse sorghum cultivars of Ethiopian origin processed into various traditional food products consumed in that country. Specifically, the objective was to determine the dynamism of in-vitro protein digestibility (IVPD) and anti-nutritional compounds across various food products prepared from diverse sorghum genotypes and their interaction between genotype and food processing methods on the bioavailability of proteins and micronutrients.

Materials and methods

Genetic materials

Fifteen sorghum cultivars and one maize (Zea mays) variety (Melkassa 1) and one tef (Eragrostis tef) variety (Boset) of Ethiopian origin were used in this study. The list and protein profile of the sorghum cultivars included in the study is presented in Table 1. Sorghum genotypes included both key landraces and released varieties in major sorghum growing areas in Ethiopia. In selecting the cultivars, efforts were made to capture as much diversity as possible not only in terms of genetics but also in their region of adaptation, physical and chemical grain attributes, and agronomic characteristics (maturity, plant height).

Table 1. Total protein, total kafirin, percent gamma kafirin profile of sorghum genotypes as related to their raw flour IPD.

Sample preparation

The food samples were prepared following the actual procedure used to by local community to prepare the different food products. Samples were prepared in the Food Science laboratory Melkassa Agricultural Research Center of the Ethiopian Institute of Agricultural Research (EIAR), at Nazareth, Ethiopia. Grains from each genotype were milled using a Cyclotec™ Cyclone sample mill (FOSS, Foss Allé 1, Hilleroed, Denmark) with a 0.5 mm sieve and the flour samples were stored in a zip-lock bags at 4°C until needed. The flour was used to prepare three different food products (fermented flat bread, unleavened bread and porridge) and a dry-cooked product was also prepared from an intact grain following procedures as described below.

Dry-cooking was performed by placing a cleaned and dried sorghum grain on a metallic pan. The pan was placed on a stove top and a medium heat was applied and this continued until the grain was cooked.

For the fermented flatbread, 200g of flour from each variety was mixed with 180 mL of water at room temperature (25°C) and kneaded for 5 min. Then a (5% on flour weight basis) pre-prepared starter yeast culture containing sufficient amount of water was poured (ten mL for each sample) on to small wells made on the surface the dough. The dough was left for 48 h to ferment at room temperature (25°C). Approximately 80g of fermented dough was mixed with 30 mL water at room temperature (25°C) followed by 200 mL boiling water and then mixed thoroughly for 1 min. The mixture was left at room temperature (25°C) until the temperature dropped to 45°C and the mixture was added back on the fermenting dough and mixed well. To this, 100 ml of water was added and the mixture was let to ferment for 3 hours at room temperature (25°C) until a foamy slurry was formed. The slurry was poured on a temperature-controlled plate pre-heated to 400°C in circular motions and covered to cook for 2 minutes.

Porridge was prepared by adding 125g of flour to 250 mL of clean boiling water followed by continuous heating at 96°C with stirring for 10 mins.

The unleavened flatbread was made in similar way as the leavened bread but not fermented. The flour was mixed with water at room temperature (25°C) and kneaded while adding water until a soft, non-sticky dough was formed. The dough was poured again in circular manner on to pre-heated clay griddle and cooked at 140°C for 10 min until a brown color was developed. All products were dried in an oven at 40°C for 24 hrs. The dry products were ground and packed in plastic zip-lock bags and along with the raw flour samples of all sorghum varieties as well as maize and tef were shipped to the USA for chemical analysis.

Sample characterization

For all tests, ground samples of four different food products, fermented flatbread, unleavened flatbread, the dry-cooked product, porridge and the raw flour was analyzed in duplicate.

In-vitro protein digestibility (IVPD) was determined according to the pepsin digestibility method [47,48]. Protein content of undigested flour and digested residue was determined through nitrogen combustion using a LECO 628 Nitrogen Determinator (LECO Corporation, St. Joseph, MI, USA) according to AACCI method 46–30.01 [49] with a N to protein conversion factor of 6.25.

Tannin concentration of the raw flour and food samples was determined using HCl vannillin assay [50]. Measurement of phytic acid was done using a commercial colorimetric kit (Megazyme phytic acid assay kit, Megazyme, Bray Business Park, Bray, Co. Wicklow, Ireland). Intrinsic phytase levels in food samples were measured using an acidic phytase activity assay using direct incubation [51]. Total Fe and Zn of the samples were analyzed using nitric-perchloric acid digestion method [52]. Analysis for Fe and Zn was carried out using an Inductively Coupled Plasma (ICP) Spectrometer (Model 720-ES ICP, Optical Emission Spectrometer, Varian Australia Pty Ltd, Mulgrave, Vic Australia).

Trypsin inhibitor activity in food samples was analyzed as described previously [53]. Preliminary tests were conducted to evaluate the need to defat samples prior to analysis and samples were defatted using three one hour extractions with hexane (1g sorghum to 10 mL of hexane) followed by air drying. No significant differences in trypsin inhibitory activity between defatted and non-defatted samples was noted, so non-defatted material was used for routine analysis. Trypsin inhibitors were extracted using one gram of sorghum flour stirred with 15 ml of 100 mM sodium phosphate buffer (pH 7.6) for 4 hrs. The slurry was centrifuged at 1789 xg for 10 min under cold conditions (0–5°C). An aliquot of 250 uL of the supernatant was dialyzed against 100 mM sodium phosphate buffer (pH 7.6) using Tube-O-DIALYZER™ Micro 4kDa MWCO (Cat# 786–611, G-bioscience, St. Louis, MO, USA) following manufacturer’s instructions. Trypsin inhibitor activity was determined as mg of trypsin inhibited per gram of flour sample (mg g-1). Raw flour samples were analyzed for kafirin content and composition using RP-HPLC with a C18 column [54].

Statistical analysis

Data were analyzed considering a 5x17 factorial arrangement in a randomized complete block design using mix model procedure, MIXED in SAS version 9.4 [55], where genotypes were treated as blocks and food processing methods as treatments. Replicates were treated as a random effect. Significant least square means were separated using Fisher’s protected least significant difference (LSD) test with the probability of type-I error (α) set at 0.05. The same α level was used to generate the Pearson correlation coefficients between measured parameters using GGally package (Schloerke, 2016) in R [56].


Protein and total kafirin profile of the 15 sorghum cultivars included in this study is presented in Table 1. The protein content among cultivars ranged from a low of 9.8% in one of the farmers’ variety Masugi-Yellow, to a high of 15.2% in the high tannin red variety, Seredo. The high lysine cultivar, Wetet Be-gunchie, has an intermediate protein content of 13.1%. Similarly, the cultivars also displayed wide variability for total kafirin with the lowest levels in variety Chiro and Wotet Be-gunchie and the highest in Teshale and Melkam. Likewise, the γ-kafirin content among cultivars ranged from 4.1% in Wetet Be-gunchie to 7.5% in variety Teshale which also had the highest total kafirin. Wetet Be-gunchie despite its intermediate protein content had the lowest amount of both total kafirin and γ-kafirin indicating that it carries significant portion of non-kafirin protein.

The analysis of variance for IVPD, micronutrient concentration and anti-nutritional compounds is presented in Table 2. The results showed significant (P ≤ 0.001) genotype, food product and genotype × food product interaction effects for IVPD, phytic acid and phytase activity. The effects of genotype and food product were also significant for trypsin inhibitor activity (TIA), and Fe and Zn concentration showing that both genotype and processing methods can alter bioavailability of nutrients from sorghum grains and food products. There was not significant genotype × food product interaction effect for TIA as well as Fe and Zn content.

Table 2. Analysis of variance of in-vitro protein digestibility (IVPD), anti-nutritional compounds and micronutrient concentration as affected by genotypes and food processing methods.

Across genotypes, the mean IVPD from raw flour samples ranged from a low of 27.7% in the high tannin variety Seredo to 75.1% in the high lysine sorghum Wetet Be-gunchie, while maize and tef had 55.8 and 62.3%, respectively (Tables 1 and 3). Sorghum cultivars including Masugi- Yellow, Degalit-Yellow, Chiro and Jigurte had a raw sample IVPD scores comparable to that of maize indicating the potential for developing high digestible sorghum varieties with comparable feed value with to that of maize. However, the results show different picture when the samples were cooked. Sorghum IVPD decreased in food products with dry cooked food having relatively higher IVPD followed by fermented bread, and the unfermented wet-cooked products having the least IVPD (Table 3). However, the reduction across genotypes was not consistent in various food products though the high tannin Seredo was consistently shown to have the least IVPD in all food products. The high lysine Wetet Be-gunchie did not maintain higher IVPD in all food products. In dry cooked samples, varieties Chiro, Teshale and the high tannin IS9302 had higher IVPD of about 30% which was markedly lower than ~40% in maize. While Meko had higher IVPD of 24% in fermented bread which is again considerably less than 45.4 and 33.8% in maize and tef, respectively. The IVPD in the unfermented food products, porridge and unleavened bread, was higher in the high lysine Wetet Be-gunchie (26.6 and 19.1%, respectively) than tef and other sorghum genotypes but lower than that of maize.

Table 3. Genotype mean estimates for in-vitro protein digestibility (IVPD) assay resulted for raw flour and four different food products.

In general, IVPD was reduced in processed foods with the degree of the reduction dependent on genotype. Table 4 shows the IVPD of cooked products as percentage of the raw flour. Reduction in IVPD of cooked products was common for all crop types with the effect being less in maize and tef compared to sorghum. Nevertheless, the extent of the reduction was markedly different between sorghum genotypes and food products. Accordingly, the IVPD in IS9302 and Teshale were least affected by dry cooking that they maintained 82 and 77.3%, respectively, of the IVPD of the raw flour. Whereas, the high lysine cultivar Wetet Be-gunchie and Degalit-Yellow maintained only 39 and 38.3%, respectively, of the digestibility of the raw flour. The relative IVPD of fermented bread among genotypes was lower but had similar trend with that of dry cooked product with IS9302 and Meko showing the least reduction by maintaining 60.7 and 60% IVPD of the raw flour, respectively, while again Wetet Be-gunchie and Massugi Yellow maintaining only 23.7 and 25.0%, respectively, of the IVPD of the raw sample. The difference among genotypes of the relative IVPD in unfermented products was relatively small compared to that of fermented bread and dry cooked product with cultivar Teshale still having the highest (40%) relative IVPD in porridge while Degalit-Yellow and Jigurte having the least, 19.9 and 18.9%, respectively. Similarly, IESV92021-DL and Melekam had relative IVPD of 35.9 and 34.9%, respectively while Masugi-Yellow and Chiro had 22.4 and 19.6% relative IVPD, respectively. In general, the data showed not only the significant effect of processing but also of the genotypes where their ranks changed according to food products. Another important observation is that digestibility of cooked products are not consistent with that of raw samples begging for the need to evaluate genotypes in cooked samples when the target is to enhance the crop as human food. Moreover, though generally lower, the difference among genotypes for IVPD of cooked samples is not as wide as in the raw flour except for few high tannin sorghums that continue to be lower. Given the little correlation between raw flour and cooked product IVPD observed in this and previous studies, and the strong association between protein content and IVPD, effort to enhance protein nutrition from sorghum based food products needs to pay as much emphasis to protein content as digestibility. In this study the low digestible sorghums such as Teshale, Melkam or 76T1#23 that had higher protein content but relatively lower raw flour IVPD may not be inferior in total protein bioavailability both in cooked and uncooked state. In other words, because IVPD is expressed as percentage it does not provide the amount of protein that become available up on digestion that bioavailability of proteins from high and low digestible sorghums may not be that different when protein content is accounted for. Actual protein availability predicted from the IVPD and protein content data for different cooked products show little difference between the high and low digestible sorghums.

Table 4. In-vitro protein digestibility of processed food products from the test genotypes expressed as percent of raw flour IVPD.

A bar chart comparing the IVPD of two representative sorghum types, the normal and high lysine line, with that of maize and tef in different food products is presented in Fig 1. In all cases cooking had negative effect on IVPD in all crops, particularly tef and sorghum, but the effect clearly varies between food products. Maize was least affected by processing methods with IVPD in raw maize sample not significantly different from the fermented bread but lower in other food products (Fig 1). However, this was different in tef and sorghum where cooking had significantly reduced IVPD in all food products with the effect being more pronounced in sorghum. This shows that sorghum is not unique in the resistance of its protein to enzymatic degradation up on cooking. Additional understanding of the key factors responsible for the decrease in protein digestibility apart from disulfide crosslinking in the γ-kafirin fraction should be the subject of future investigation.

Fig 1. In-vitro protein digestibility (IVPD) across food products of maize, tef and high-lysine and average of 14 normal sorghum genotypes.

Bars in each panel followed by the same letter are not significantly different.

Contour plots based on five genotypes with the highest and lowest IVPD were created in order to provide a graphical view of how IVPD varied between genotypes and food products (Fig 2). It was evident from the plot that the IVPD score of genotypes was specific to the food products. While the high-lysine sorghum genotype Wetet Be-gunchie had the highest IVPD for raw flour, porridge and unleavened bread, the genotype Meko was the highest for fermented flatbread and genotype Chiro for the dry cooked product. The high tannin genotype Seredo, however, had the lowest IVPD in all food products with porridge and unleavened bread having the least IVPD.

Fig 2.

Contour plots showing 10 sorghum genotypes with the highest (red dots) and lowest (yellow dots) in-vitro protein digestibility in (a) raw flour, (b) fermented flatbread, (c) porridge, (d) unleavened flatbread and, (e) dry-cooked sorghum. Axis values in each plot depict protein digestibility range observed for each cooked product and raw flour.

The Venn diagram presented in Fig 3 shows the pattern of overlap for IVPD of the top five (Fig 3A) and bottom five (Fig 3B) genotypes tested in different food products. None of the cultivars consistently expressed high IVPD in all food products. However, many of the genotypes that had the highest raw flour IVPD were also among the highest in one or more of the food products. Cultivar Chiro was among the top five in three food products and was also among the highest in the raw flour. Masugi-Yellow, Meko and Jigurte were good only in one product. The other seven varieties were among the top in two products with five of them good for making fermented bread. On the other hand, all of the cultivars that were among the top five IVPD group for raw flour also had among the highest IVPD in one or more cooked products. Similarly, all cultivars that were among the lowest IVPD group in raw flour were also among the lowest IVPD group in one or more cooked products (Fig 3A and 3B and Table 3). Further, all of the cultivars that were among the highest IVPD group, except Wetet Be-gunchie, in one or more cooked products were also among the lowest IVPD group in one or more of the other cooked products. Whereas three of the lowest raw flour IVPD cultivars that were among the lowest IVPD group in one or more of cooked products were not among the highest IVPD group in any of the cooked products. Only Seredo was consistently grouped among the low IVPD category in all cooked products as well as the raw flour.

Fig 3.

Venn diagrams showing the distribution of sorghum genotypes with (a) highest and, (b) lowest IVPD score for cooked products. The underlined cultivars are those highest or lowest IVPD score in raw flour. Tannin containing genotypes are marked with a star symbol.

As expected, there were positive correlations between total protein and total kafirin (r2 = 0.72) as well as total protein and γ-kafirin (r2 = 0.44) (Fig 4). Correlations of raw flour IVPD with total protein (r2 = 0.38), total kafirin (r2 = 0.66) and γ-kafirin (r2 = 0.52) were significant and negative as was IVPD of unleavened flatbread and porridge with total kafirin (r2 = 0.18 and r2 = 0.14, respectively). Raw flour IVPD showed positive correlation with porridge IVPD (r2 = 0.48) and unleavened flatbread (r2 = 0.49). Supporting this result was a positive correlation between IVPD of porridge and unleavened flatbread (r2 = 0.45). Likewise, IVPD in fermented bread and dry-cooked product also showed a positive correlation (r2 = 0.18).

Fig 4. Correlation matrix for total protein content, total kafirin content, gamma-kafirin content and IVPD broken down based on food products and raw flour.

Statistical significance at P ≤ 0.05, 0.01 and 0.001, are shown using *, **, ***, respectively.

Other grain characteristics may be related to IVPD, thus we evaluated the effects of phytic acid, Fe and Zn concentrations and the levels of phytase and trypsin inhibitor in the raw flour and cooked food samples on IVPD. Table 5 shows the estimate of these parameters in all product samples tested. As revealed by the data, mean phytic acid across cultivars was highest in the raw flour followed by the unfermented products (porridge and unleavened bread), and lowest in the fermented bread while the intrinsic phytase activity was highest in fermented bread and lowest in porridge and unleavened bread. Trypsin inhibitor activity was significantly lower in raw flour and increased during food processing. Fe and Zn contents showed considerable variability among food products while phytate/Fe and phytate/Zn molar ratios showed a consistent pattern with raw flour > porridge > unleavened flatbread > dry-cooked > fermented flatbread (Table 5). Fermentation appeared to have positive effect on enhancing protein nutrition in that it considerably reduced phytic acid and TIA, and significantly improved phytase activity (Table 5).

Table 5. Across genotype mean estimates for all measured parameters and phytate/mineral molar ratios in raw flour and four different food products.

Fig 5 displays the profile of anti-nutritional factors and intrinsic phytase levels in sorghum cultivars shown to have high IVPD in two or more cooked products as well as raw flour. Phytic acid concentration was consistently and significantly higher in Wetet Be-gunchie in both raw flour and all of the cooked products, with the levels slightly lower in the fermented flat bread (Fig 5A). The two other cultivars, Degalit-Yellow and Chiro not only had markedly lower phytic acid concentration than Wetet Be-gunchie but also were different from each other with Degalit-Yellow having higher concentration than Chiro. Wetet Be-gunchie also showed higher levels of phytase activity in all products except in the fermented flat bread and dry-cooked product where the difference was not significant (Fig 5B). Again Degalit-Yellow had significantly higher phytase levels than Chiro in raw flour as well as in unfermented food products (porridge and unleavened bread). Trypsin inhibitor on the other hand was not significant between the three genotypes except in the cooked product porridge where Wetet Be-gunchie was shown to be significantly lower than the other genotypes (Fig 5C). In general, phytic acid concentration showed similar variability between genotypes in all food products with Wetet Be-gunchie containing the highest concentration (Fig 5A). Phytase activity was still higher in Wetet Be-gunchie but significant only in raw flour and unfermented products (Fig 5B). Trypsin inhibitor, however, was not significantly different except in porridge where Wetet Be-gunchie had the lowest activity (Fig 5C).

Fig 5.

Levels of anti-nutritional factors and intrinsic phytase levels in three sorghum cultivars shown to have among the highest IVPD in raw flour as well as in three or more cooked products: (a) phytic acid concentration, (b) phytase concentration and, (c) trypsin inhibitor levels.

Fig 6 shows the correlations between anti-nutritional factors, phytase activity, mineral concentrations and phytate/mineral molar ratios. While the raw flour IVPD showed significant and positive correlation with phytic acid (r2 = 0.03), trypsin inhibitory activity showed a significant and negative correlation with IVPD (r2 = 0.11). Correlations of phytase with Fe (r2 = 0.10) and Zn content (r2 = 0.06), and Fe with Zn (r2 = 0.10) were positive. Moreover, phytic acid showed a negative correlation with trypsin inhibitory activity (r2 = 0.05). Phytate/Fe and phytate/Zn molar ratios, on the other hand, showed positive correlation with IVPD (r2 = 0.21 and r2 = 0.10, respectively) and phytic acid (r2 = 0.25 and r2 = 0.02, respectively) and negative correlation with phytase (r2 = 0.16 and r2 = 0.24, respectively).

Fig 6. Correlation between IVPD, anti-nutritional factors (phytic acid and trypsin inhibitor), phytase concentration, and micronutrient content in raw flour samples of fifteen sorghum cultivars.

*, **, ***, = statistically significant at P ≤ 0.05, 0.01 and 0.001, levels of probability, respectively.


The significance of the discoveries of the effect of sorghum protein body structure [6,8,57] and disulfide cross-linking [4,7,58] on protein digestibility have received significant research attention in the past several years. Prior to that, the discovery of the Ethiopian high lysine landraces and the later discovery of the high lysine mutants had both attracted similar interest and the research work that followed has generated significant information on the nutritional quality of sorghum [59,60,61]. An equally important topic that has not attracted as much research attention is the impact of food processing on bioavailability of sorghum proteins. While not a new topic in sorghum utilization studies, research on local food processing methods, especially in relation to its potential outcome for enhancing the nutrition of smallholder households dependent on the crop. The goal of the current study was not to elucidate the mechanisms of how food processing impacts sorghum biomolecules, but rather to demonstrate the impacts of traditional food processing methods by native consumers of the crop on availability of protein and micronutrients. An additional outcome of this research was to identify areas where the science of food chemistry can be applied to add value to these processes and improve nutritional value of sorghum-based diets.

IVPD is genotype and food product specific

As reported in several previous findings [47,62,63], the IVPD of sorghum significantly reduces when cooked to various food products and this was true for every genotype (Table 2). This was also true for tef as well confirming previous findings [9,47,62] while the reduction in IVPD of maize cooked products was substantially lower (Fig 1). The mean IVPD across genotypes was reduced in all food products compared to the raw flour. But the extent of the reduction was different for different food products with the mean IVPD ranking of raw > dry-cooked > fermented bread > porridge and unleavened bread and agrees with recently reported results [64]. Earlier studies have reported popping or roasting to have no negative effect on sorghum IVPD [43] but in this study it did affect IVPD in all sorghum varieties as well as in maize genotype. Several other investigations have shown fermentation to increase the IVPD of sorghum [39,6567] compared to unfermented wet-cooking such as in unleavened bread and porridge which severely reduces digestibility [9,62,63] as is also the case in the current study.

However, IVPD showed unpredictable fluctuations when the flour was processed into various food types (Table 1 and Fig 2) even for genotypes that belonged to either of a known high or low uncooked IVPD groups (Figs 2 and 3). In other words, the pattern of the reduction in IVPD across genotypes for different food products was not consistent and rarely depended on that of raw flour IVPD. The only exception was the high tannin genotypes that had consistently lower IVPD in all food products as well as the raw sample. At the same time there was significant differences in IVPD of food products made from different genotypes but the differences did not show a consistent pattern. This could be explained by the significant genotype × food type interaction (P <0.001) observed for IVPD (Table 1).

The best demonstration of the observed genotype × food type interaction effect on IVPD was that of the high-lysine sorghum genotype Wetet Be-gunchie which had the highest IVPD for raw flour, porridge and unleavened bread, all of which are wet cooked products. However, this genotype did not fall among the highest IVPD for fermented flatbread and dry-cooked product (Table 2 and Fig 2). Other genotypes showed similar variation but not necessarily the same pattern. This result depicts a very important characteristics of sorghum in general where cooking not only decreased IVPD significantly but also unpredictably. The difference in IVPD of raw flour and cooked products has complicated the cultivar improvement process rendering the standard raw flour digestibility results irrelevant for food value. Added to this complication was the unpredictable difference in IVPD of various food products made from different varieties. This result explains why traditional farmers sort varieties not just by their grain color or agronomic value but also by the quality of food products they make. In Ethiopia, an ancient country where sorghum was believed to have originated, local farmers categorize the crop based on its utilization traits; they have sorghum types specifically used for making popped snacks, others for making bread and still others for local brew giving them distinct names that describes their use. This and previous research results suggest that sorghum improvement efforts for food use needs to take into account the intended use of the crop and conduct tests and selections for specific food products. Since food types and products vary from place to place depending on the culture of the people, the process of cultivar development may become complicated due to varying needs of communities.

The other important result was the association between IVPD and the content of grain protein and protein fractions. There was a significant negative correlation between IVPD and total protein, total kafirin as well as gamma kafirin content in raw flour samples (Fig 4). The negative association between γ-kafirin and IVPD agrees with the previous reports that γ-kafirin is resistant to pepsin owing to its high cysteine content that leads to disulphide cross-links [63,68,69]. The fact that Wetet Be-gunchie, the high lysine genotype with high protein digestibility also had low γ-kafirin agrees with the previous reports.

However, when the flour was cooked to different food products, the strength of the association between IVPD and protein/kafirin content also markedly reduced with the association becoming non-existent in fermented breads. It is well known that processing flour into different food products alters the structure of the product affecting enzymatic activity towards the proteins. The standard IVPD assay determines only the percentage of the grain protein content digested and does not account for the actual amount of digested protein. When the amount digested is considered, the amount of protein digested per unit of flour/food sample was not very different between genotypes except for the high lysine Wetet Be-gunchie and the high tannin Seredo in raw flours. For example, 100g of flour from variety Melekam (15% protein content) at the reported IVPD of 38% yields 5.7g protein digested. A low protein (10%) cultivar such as Masugi-Yellow with a reported IVPD of 61% has a yield of 6.1g protein digested, not very different from the low digestible Melekam with the higher total protein. This was the case for all other varieties except the two extremes, Wetet Be-gunchie and Seredo. The uniqueness with these two genotypes may have to do with their known properties. Wetet Be-gunchie has intermediate protein content (13.8%) but low in total kafirin unlike other genotypes of moderately high protein content indicating that the cultivar has relatively higher proportion non-kafirin proteins that are high digestible. Moreover, this genotype also has relatively low γ-kafirin content, a portion of sorghum kafirin known to be resistant to enzymatic degradation. On the other hand, the high tannin genotype Seredo is high protein and high kafrin type, the only difference from other genotypes is its high tannin content that perhaps compromised its IVPD both in cooked and uncooked state.

While the strong negative correlation observed between γ-kafirin and IVPD in the raw flour (Fig 4) and a reduction in cooked products may be expected given the reported effect of γ-kafirin [10,70], the complete absence of such association in fermented food product was not clear. The degradation of starch granules and protein during fermentation may increase access for pepsin to more effectively digest protein structures and that the levels of γ-kafirin in raw sample becomes less relevant. The absolute reduction in IVPD of fermented product, however, has to do with the cooking of the protein that may have resulted in altered protein structures through disulfide cross linking or increased hydrophobicity [58,71].

In general, processing cereals into food products reduces the availability of proteins, especially in sorghum and the extent of the reduction depends on both genotype and the type of food product. Even the known high digestible Ethiopian high lysine landrace was not much different from other cultivars when cooked (Table 3) and the reduction has no specific pattern (Table 2) although unfermented wet cooked products have relatively lower IVPD.

Although all genotypes have undergone reduction in IVPD when cooked, the extent to which the reduction occurred was different between genotypes as well as food products resulting in significant genotype by process interaction. The interactions, however, did not seem to follow specific pattern (Table 3). Certain genotypes were better for making fermented bread while others were preferable for making other products. Thus future improvement of sorghum needs to identify grain characteristics that makes them suitable for producing specific food products and direct selection for particular end use traits. However, IVPD is just one trait and final product quality and consumer acceptance also needs to be considered. Moreover, as pointed out elsewhere, IVPD assays do not directly measure the actual amount of protein that becomes available after digestion has taken place (i.e. total digested protein). Another important factor is the difference in gastric emptying time for different foods. Sorghum foods have been reported to have longer stomach emptying time than foods made from other grains [72]. With a longer time in the stomach, proteins from sorghum food may have extended period of exposure to digestive enzymes that the standard pepsin assay may not accurately mimic. If this is confirmed, then protein digestibility may stop to be burning issue in protein nutrition and protein content may become as important if not more. In the current study, the total amount of protein that becomes available after digestion was not different among genotypes especially for cooked products. Hence, in addition to IVPD, attention may be paid to protein content.

Several previous studies have demonstrated considerable differences in percent protein digestibility of raw samples [6,8,9,62,66,69] and industry processed foods [35,37,64,67,7376]. The current study was also based on raw germplasm samples and food products but deviates greatly from tailored food processing procedures [64, 76]. Many studies that have investigated IVPD of cooked sorghum have used a single 10 or 20 min boiling step of raw flour suspended in water to mimic cooking of porridge [47,62,63,69,77,78]. The current investigation was based on actual foods produced as consumed by local communities in Ethiopia rather than laboratory procedures designed to mimic food production.

Anti-nutritional factors and mineral content

There was significant effect of genotype and food product effect on anti-nutritional factors (Table 2). Phytic acid content in different food products ranged from 0.62 in fermented bread to 0.77 in raw flour all within the previously reported range of 0.4 and 3.5 g 100g-1 [79,80,81]. Food processing seem to have slightly reduced phytic acid levels with fermentation and dry-cooking significantly reduced the compound presumably due to increased phytase activity in fermented and dry-cooked products (Table 5). This agrees with previous studies that showed the effects of cooking and fermentation on phytic acid levels [82]. However, the positive effects of cooking on phytic acid were compromised by increased TIA in all products which along with other chemical changes to the protein body and starch-protein complex may have undermined the overall digestibility in cooked products (Table 3). Adding to this complexity, certain genotypes such as Wetet Be-gunchie that have high raw flour digestibility also have higher phytic acid content in all food products (Fig 5). They also have higher phytase activity especially in raw and unfermented wet-cooked products where IVPD from these genotype was higher (Table 3). While TIA was largely not significantly different among high digestible cultivars (Fig 5) certain low digestible varieties such as Seredo, or medium digestible ones such as Teshale also have similar TIA with that of Wetet Be-gunchie and other high digestible cultivars (S1 Table). These all add to the complexity of the sorghum protein digestibility trait.

Similarly, Fe and Zn content across food products was significantly different among genotypes (S2 Table) with cooking appearing to have effect on both minerals (Table 5). However, not all genotypes were affected by cooking with Fe being significantly different between food products for all genotypes except Dagim, Masugi-Yellow, IS9302, Jigurte and Seredo while Zn content significant only for seven of the fifteen genotypes (S2 Table). Cooking seem to have markedly increased total Fe content with fermentation having the greatest effect (Table 5). While the cause of this increase is not known for certain, it may be the result of the degradation of phytic acid in cooked products and thus release of minerals chelated it. However, the extent of reduction in phytic acid in unfermented food products was very low compared to the increase in total Fe in these products. Nevertheless, similar results have been reported by previous investigators especially of that of fermented products [80,83]. While bioavailability of these minerals was not directly measured in the current project, past studies have used phytate/mineral ratio as potential indicator for bioavailability of both Fe and Zn [42,80]. Using this indicator, bioavailability of Fe and Zn across food products evaluated in the current study ranged from 10–17.9 and 25.1–35.4, respectively (Table 5) with the range for Zn differing slightly from results reported earlier [42]. Due to its high density of negatively charged phosphate groups, phytate binds with minerals and inhibit their availability, especially Fe and Zn bioavailability have been reported to be highly compromised by phytic acid [8489]. Likewise, a reduction in phytate content improves the availability of these minerals [88,90,91]. Based on the phytate/mineral ratio in the current study, the trend in the availability of Fe and Zn follows the order of fermented bread > dry cooked > wet-cooked > raw flour. Hence, food processing in addition to its effect on IVPD, can influence the availability of essential nutrients and this is of significant practical value to smallholder farmers.


The study showed that both genotype, food processing method and their intercation have highly significant effect on IVPD of sorghum traditional foods commonly used by rural communities in Ethiopia. The findings shed light on possible avenues towards improving IVPD through manipulating food processing techniques and selection of specific germplasm for making specific food product. Trypsin inhibitor in most of the genotypes was stable and resistant to degradation by food processing. However, processing decreased phytic acid concentration making both protein and minerals more bioavailable.

Supporting information

S1 Table. Trypsin inhibitor activity of genotypes across food processing methods.


S2 Table. Iron and Zinc content of genotypes across food processing methods.



This research was supported by the American People through the United States Agency for International Development (USAID) Feed the Future Innovation Lab for Collaborative Research on Sorghum and Millet (SMIL) under Cooperative Agreement No. AID-OAA-A-13-00047.


  1. 1. Rooney LW, Pflugfelder RL. Factors Affecting Starch Digestibility with Special Emphasis on Sorghum and Corn 1. Journal of Animal Science. 1986 Nov 1;63(5):1607–23. pmid:3539904
  2. 2. Axtell JD, Kirleis AW, Hassen MM, Mason ND, Mertz ET, Munck L. Digestibility of sorghum proteins. Proceedings of the National Academy of Sciences. 1981 Mar 1;78(3):1333–5.
  3. 3. Axtell JD, Ejeta G, Munck L. Sorghum nutritional quality-progress and prospects. SORGHUM. 1982 Oct:589.
  4. 4. Hamaker BR, Kirleis AW, Butler LG, Axtell JD, Mertz ET. Improving the in vitro protein digestibility of sorghum with reducing agents. Proceedings of the National Academy of Sciences. 1987 Feb 1;84(3):626–8.
  5. 5. Tanksley TD, Knabe DA. Ileal digestibilities of amino acids in pig feeds and their use in formulating diets. ed. 2.
  6. 6. Duodu KG, Taylor JR, Belton PS, Hamaker BR. Factors affecting sorghum protein digestibility. Journal of Cereal Science. 2003 Sep 1;38(2):117–31.
  7. 7. Wong JH, Lau T, Cai N, Singh J, Pedersen JF, Vensel WH, et al. Digestibility of protein and starch from sorghum (Sorghum bicolor) is linked to biochemical and structural features of grain endosperm. Journal of cereal science. 2009 Jan 1;49(1):73–82.
  8. 8. Oria MP, Hamaker BR, Axtell JD, Huang CP. A highly digestible sorghum mutant cultivar exhibits a unique folded structure of endosperm protein bodies. Proceedings of the National Academy of Sciences. 2000 May 9;97(10):5065–70.
  9. 9. Duodu KG, Nunes A, Delgadillo I, Parker ML, Mills EN, Belton PS, et al. Effect of grain structure and cooking on sorghum and maize in vitro protein digestibility. Journal of Cereal Science. 2002 Feb 1;35(2):161–74.
  10. 10. Da Silva LS, Jung R, Zhao ZY, Glassman K, Taylor J, Taylor JR. Effect of suppressing the synthesis of different kafirin sub-classes on grain endosperm texture, protein body structure and protein nutritional quality in improved sorghum lines. Journal of cereal science. 2011 Jul 1;54(1):160–7.
  11. 11. Awika JM, Rooney LW. Sorghum phytochemicals and their potential impact on human health. Phytochemistry. 2004 May 1;65(9):1199–221. pmid:15184005
  12. 12. Cheryan M, Rackis JJ. Phytic acid interactions in food systems. Critical Reviews in Food Science & Nutrition. 1980 Dec 1;13(4):297–335.
  13. 13. Graf E, Eaton JW. Antioxidant functions of phytic acid. Free Radical Biology and Medicine. 1990 Jan 1;8(1):61–9. pmid:2182395
  14. 14. Ali N, Paul S, Gayen D, Sarkar SN, Datta K, Datta SK. Development of low phytate rice by RNAi mediated seed-specific silencing of inositol 1, 3, 4, 5, 6-pentakisphosphate 2-kinase gene (IPK1). PloS one. 2013 Jul 2;8(7):e68161. pmid:23844166
  15. 15. Hunt JR. Bioavailability of iron, zinc, and other trace minerals from vegetarian diets. The American journal of clinical nutrition. 2003 Sep 1;78(3):633S–9S. pmid:12936958
  16. 16. Hurrell RF, Reddy MB, Juillerat MA, Cook JD. Degradation of phytic acid in cereal porridges improves iron absorption by human subjects. The American journal of clinical nutrition. 2003 May 1;77(5):1213–9. pmid:12716674
  17. 17. Hurrell R, Egli I. Iron bioavailability and dietary reference values–. The American journal of clinical nutrition. 2010 Mar 3;91(5):1461S–7S. pmid:20200263
  18. 18. Kakade ML, Simons N, Liener IE. Evaluation of natural vs. synthetic substrates for measuring the antitryptic activity of soybean samples. Cereal chemistry. 1969.
  19. 19. Savelkoul FH, Van der Poel AF, Tamminga S. The presence and inactivation of trypsin inhibitors, tannins, lectins and amylase inhibitors in legume seeds during germination. A review. Plant Foods for Human Nutrition. 1992 Jan 1;42(1):71–85. pmid:1372122
  20. 20. Song HK, Suh SW. Kunitz-type soybean trypsin inhibitor revisited: refined structure of its complex with porcine trypsin reveals an insight into the interaction between a homologous inhibitor from Erythrina caffra and tissue-type plasminogen activator1. Journal of molecular biology. 1998 Jan 16;275(2):347–63.
  21. 21. Guillamon E, Pedrosa MM, Burbano C, Cuadrado C, de Cortes Sánchez M, Muzquiz M. The trypsin inhibitors present in seed of different grain legume species and cultivar. Food Chemistry. 2008 Mar 1;107(1):68–74.
  22. 22. Haynes R, Feeney RE. Fractionation and properties of trypsin and chymotrypsin inhibitors from lima beans. Journal of Biological Chemistry. 1967 Nov 25;242(22):5378–85. pmid:6070853
  23. 23. Feeney RE, Means GE, Bigler JC. Inhibition of human trypsin, plasmin, and thrombin by naturally occurring inhibitors of proteolytic enzymes. Journal of Biological Chemistry. 1969 Apr 25;244(8):1957–60. pmid:4238526
  24. 24. Tan CG, Stevens FC. Amino acid sequence of lima bean protease inhibitor component IV. The FEBS Journal. 1971 Feb 1;18(4):503–14.
  25. 25. Melville JC, Scandalios JG. Maize endopeptidase: genetic control, chemical characterization, and relationship to an endogenous trypsin inhibitor. Biochemical genetics. 1972 Aug 1;7(1):15–pmid:4557513
  26. 26. Wen L, Huang JK, Zen KC, Johnson BH, Muthukrishnan S, MacKay V, et al. Nucleotide sequence of a cDNA clone that encodes the maize inhibitor of trypsin and activated Hageman factor. Plant molecular biology. 1992 Feb 1;18(4):813–4. pmid:1558956
  27. 27. Mikola JU, Kirsi M. Differences between endospermal and embryonal trypsin inhibitors in barley, wheat, and rye. Acta Chem. Scand. 1972 Jan 1;26(2):787–95.
  28. 28. Mikola J, Suolinna EM. Purification and properties of a trypsin inhibitor from barley. The FEBS Journal. 1969 Jul 1;9(4):555–60.
  29. 29. Odani S, Koide T, Ono T. The complete amino acid sequence of barley trypsin inhibitor. Journal of Biological Chemistry. 1983 Jul 10;258(13):7998–8003. pmid:6345537
  30. 30. Mikola M, Mikkonen A. Occurrence and stabilities of oat trypsin and chymotrypsin inhibitors. Journal of cereal science. 1999 Nov 1;30(3):227–35.
  31. 31. Tashiro M. Partial purification and some properties of a trypsin inhibitor from rice bran. Agricultural and Biological Chemistry. 1978;42(6):1119–24.
  32. 32. Wang X, Hu L, Zhou G, Cheng J, Lou Y. Salicylic acid and ethylene signaling pathways are involved in production of rice trypsin proteinase inhibitors induced by the leaf folder Cnaphalocrocis medinalis (Guenée). Chinese science bulletin. 2011 Aug 1;56(22):2351–8.
  33. 33. José Filho XA. Trypsin inhibitors in sorghum grain. Journal of Food Science. 1974 Mar 1;39(2):422–3.
  34. 34. Kumar PM, Virupaksha TK, Vithayathil PJ. Sorghum proteinase inhibitors. Chemical Biology & Drug Design. 1978 Oct 1;12(4):185–96.
  35. 35. Graham GG, MacLean WC Jr, Morales E, Hamaker BR, Kirleis AW, Mertz ET, et al. Digestibility and utilization of protein and energy from Nasha, a traditional Sudanese fermented sorghum weaning food. The Journal of nutrition. 1986 Jun 1;116(6):978–84. pmid:3088228
  36. 36. Mukuru SZ, Butler LG, Rogler JC, Kirleis AW, Ejeta G, Axtell JD, et al. Traditional processing of high-tannin sorghum grain in Uganda and its effect on tannin, protein digestibility, and rat growth. Journal of agricultural and food chemistry. 1992 Jul;40(7):1172–5.
  37. 37. Taylor J, Taylor J. Alleviation of the adverse effect of cooking on sorghum protein digestibility through fermentation in traditional African porridges. International journal of food science & technology. 2002 Feb 1;37(2):129–37.
  38. 38. Makokha AO, Oniang'o R, Njoroge SM, Kinyanjui PK. Effect of malting on protein digestibility of some sorghum (Sorghum bicolor) varieties grown in Kenya. African Journal of Food, Agriculture, Nutrition and Development. 2002;2(2):59–66.
  39. 39. Osman MA. Changes in sorghum enzyme inhibitors, phytic acid, tannins and in vitro protein digestibility occurring during Khamir (local bread) fermentation. Food Chemistry. 2004 Nov 1;88(1):129–34.
  40. 40. Raihanatu MB, Modu S, Falmata AS, Shettima YA, Heman M. Effect of processing (sprouting and fermentation) of five local varieties of sorghum on some biochemical parameters. Biokemistri. 2011;23(2).
  41. 41. Kuo HT, Wu PH, Chung HH, Jiang CM, Wu MC. Koji fermentation improve the protein digestibility of sorghum. Journal of Food Processing and Preservation. 2013 Oct 1;37(5):419–23.
  42. 42. Afify AE, El-Beltagi HS, El-Salam SM, Omran AA. Protein solubility, digestibility and fractionation after germination of sorghum varieties. PLoS One. 2012 Feb 3;7(2):e31154. pmid:22319611
  43. 43. Parker ML, Grant A, Rigby NM, Belton PS, Taylor JR. Effects of popping on the endosperm cell walls of sorghum and maize. Journal of Cereal Science. 1999 Nov 1;30(3):209–16.
  44. 44. Correia I, Nunes A, Barros AS, Delgadillo I. Comparison of the effects induced by different processing methods on sorghum proteins. Journal of cereal science. 2010 Jan 1;51(1):146–51.
  45. 45. Hamaker BR, Mertz ET, Axtell JD. Effect of extrusion on sorghum kafirin solubility. Cereal chemistry. 1994 Sep 1:71:515–.
  46. 46. de Mesa-Stonestreet NJ, Alavi S, Gwirtz J. Extrusion-enzyme liquefaction as a method for producing sorghum protein concentrates. Journal of food engineering. 2012 Jan 1;108(2):365–75.
  47. 47. Mertz ET, Hassen MM, Cairns-Whittern C, Kirleis AW, Tu L, Axtell JD. Pepsin digestibility of proteins in sorghum and other major cereals. Proceedings of the National Academy of Sciences. 1984 Jan 1;81(1):1–2.
  48. 48. Cremer JE, Liu L, Bean SR, Ohm JB, Tilley M, Wilson JD et al. Impacts of kafirin allelic diversity, starch content, and protein digestibility on ethanol conversion efficiency in grain sorghum. Cereal Chemistry. 2014 May;91(3):218–27.
  49. 49. International AACC. Method 46‐30.01. Crude protein Combustion method. Approved methods of analysis. 1999.
  50. 50. Price ML, Van Scoyoc S, Butler LG. A critical evaluation of the vanillin reaction as an assay for tannin in sorghum grain. Journal of Agricultural and Food Chemistry. 1978 May;26(5):1214–8.
  51. 51. Greiner R, Egli I. Determination of the activity of acidic phytate-degrading enzymes in cereal seeds. Journal of agricultural and food chemistry. 2003 Feb 12;51(4):847–50. pmid:12568536
  52. 52. Gieseking JE, Snider HJ, Getz CA. Destruction of organic matter in plant material by the use of nitric and perchloric acids. Industrial & Engineering Chemistry Analytical Edition. 1935 May 1;7(3):185–6.
  53. 53. Smith C, Van Megen W, Twaalfhoven L, Hitchcock C. The determination of trypsin inhibitor levels in foodstuffs. Journal of the Science of Food and Agriculture. 1980 Apr 1;31(4):341–50. pmid:7190200
  54. 54. Bean SR, Ioerger BP, Blackwell DL. Separation of kafirins on surface porous reversed-phase high-performance liquid chromatography columns. Journal of agricultural and food chemistry. 2010 Dec 9;59(1):85–91. pmid:21141963
  55. 55. Institute Inc SAS. Base SAS® 9.4 procedures guide. CaryNC: SASInsti—mte Inc. 2Oo4. 2011:32–55.
  56. 56. Team RC. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2016.
  57. 57. Benmoussa M, Chandrashekar A, Ejeta G, Hamaker BR. Cellular response to the high protein digestibility/high-lysine (hdhl) sorghum mutation. Plant Science. 2015 Dec 1;241:70–7.
  58. 58. Taylor JR. Non-starch polysaccharides, protein and starch: form function and feed–highlights on sorghum. In Proceedings, Australian Poultry Science Symposium 2005 (Vol. 17, No. 9, p. 16).
  59. 59. Gebrekidan B, Kebede Y. The Traditional culture and yield potentials of the Ethiopian high Lysine sorghums. Ethiopian Journal of Agricultural Sciences. 1979.
  60. 60. Ejeta G, Hassen MM, Mertz ET. In vitro digestibility and amino acid composition of pearl millet (Pennisetum typhoides) and other cereals. Proceedings of the national academy of sciences. 1987 Sep 1;84(17):6016–9.
  61. 61. Tesso T, Hamaker BR, Ejeta G. Sorghum protein digestibility is affected by dosage of mutant alleles in endosperm cells. Plant Breeding. 2008 Dec;127(6):579–86.
  62. 62. Hamaker BR, Kirleis AW, Mertz ET, Axtell JD. Effect of cooking on the protein profiles and in vitro digestibility of sorghum and maize. Journal of Agricultural and Food Chemistry. 1986 Jul;34(4):647–9.
  63. 63. Oria MP, Hamaker BR, Shull JM. Resistance of Sorghum. alpha.-,. beta.-, and. gamma.-Kafirins to Pepsin Digestion. Journal of Agricultural and Food Chemistry. 1995 Aug;43(8):2148–53
  64. 64. Taylor J, Taylor JR. Protein biofortified sorghum: effect of processing into traditional African foods on their protein quality. Journal of agricultural and food chemistry. 2011 Feb 21;59(6):2386–92. pmid:21338110
  65. 65. Taylor JR, Novellie L, Liebenberg NV. Sorghum protein body composition and ultrastructure. Cereal Chem. 1984;6:69–73.
  66. 66. Chavan UD, Chavan JK, Kadam SS. Effect of Fermentation on Soluble Proteins and In Vitro Protein Digestibility of Sorghum, Green Gram and Sorghum‐Green Gram Blends. Journal of Food Science. 1988 Sep 1;53(5):1574–5.
  67. 67. Ibrahim FS, Babiker EE, Yousif NE, El Tinay AH. Effect of fermentation on biochemical and sensory characteristics of sorghum flour supplemented with whey protein. Food Chemistry. 2005 Sep 1;92(2):285–92.
  68. 68. Oom A, Pettersson A, Taylor JR, Stading M. Rheological properties of kafirin and zein prolamins. Journal of Cereal Science. 2008 Jan 1;47(1):109–16.
  69. 69. Weaver CA, Hamaker BR, Axtell JD. Discovery of grain sorghum germ plasm with high uncooked and cooked in vitro protein digestibilities. Cereal Chemistry. 1998 Sep;75(5):665–70.
  70. 70. Grootboom AW, Mkhonza NL, Mbambo Z, O’Kennedy MM, Da Silva LS, Taylor J, et al. Co-suppression of synthesis of major α-kafirin sub-class together with γ-kafirin-1 and γ-kafirin-2 required for substantially improved protein digestibility in transgenic sorghum. Plant cell reports. 2014 Mar 1;33(3):521–37. pmid:24442398
  71. 71. Duodu KG, Nunes A, Delgadillo I, Belton PS. Low protein digestibility of cooked sorghum—causes and needs for further research. InAFRIPRO, workshop on the proteins of sorghum and millets: Enhancing nutritional and functional properties for Africa, Pretoria, South Africa 2003.
  72. 72. Lemlioglu-Austin D, Turner ND, McDonough CM, Rooney LW. Effects of sorghum [Sorghum bicolor (L.) Moench] crude extracts on starch digestibility, estimated glycemic index (EGI), and resistant starch (RS) contents of porridges. Molecules. 2012 Sep 17;17(9):11124–38. pmid:22986923
  73. 73. MacLean WC Jr, López de Romaña G, Gastañaduy A, Graham GG. The effect of decortication and extrusion on the digestibility of sorghum by preschool children. The Journal of nutrition. 1983 Oct 1;113(10):2071–7. pmid:6413665
  74. 74. Saravanabavan SN, Shivanna MM, Bhattacharya S. Effect of popping on sorghum starch digestibility and predicted glycemic index. Journal of food science and technology. 2013 Apr 1;50(2):387–92. pmid:24425932
  75. 75. Mkandawire NL, Weier SA, Weller CL, Jackson DS, Rose DJ. Composition, in vitro digestibility, and sensory evaluation of extruded whole grain sorghum breakfast cereals. LWT-Food Science and Technology. 2015 Jun 1;62(1):662–7.
  76. 76. Fapojuwo OO, Maga JA, Jansen GR. Effect of extrusion cooking on in vitro protein digestibility of sorghum. Journal of Food Science. 1987 Jan 1;52(1):218–9.
  77. 77. Duodu KG, Tang H, Grant A, Wellner N, Belton PS, Taylor JR. FTIR and Solid State13C NMR Spectroscopy of Proteins of Wet Cooked and Popped Sorghum and Maize. Journal of Cereal Science. 2001 May 1;33(3):261–9.
  78. 78. Ezeogu LI, Duodu KG, Taylor JR. Effects of endosperm texture and cooking conditions on the in vitro starch digestibility of sorghum and maize flours. Journal of Cereal Science. 2005 Jul 1;42(1):33–44.
  79. 79. Kasim AB, Edwards HM. The analysis for inositol phosphate forms in feed ingredients. Journal of the Science of Food and Agriculture. 1998 Jan 1;76(1):1–9.
  80. 80. Kayode AP, Linnemann AR, Nout MJ, Van Boekel MA. Impact of sorghum processing on phytate, phenolic compounds and in vitro solubility of iron and zinc in thick porridges. Journal of the Science of Food and Agriculture. 2007 Apr 15;87(5):832–8.
  81. 81. Schlemmer U, Frølich W, Prieto RM, Grases F. Phytate in foods and significance for humans: food sources, intake, processing, bioavailability, protective role and analysis. Molecular nutrition & food research. 2009 Sep 1;53(S2).
  82. 82. Marfo EK, Simpson BK, Idowu JS, Oke OL. Effect of local food processing on phytate levels in cassava, cocoyam, yam, maize, sorghum, rice, cowpea, and soybean. Journal of Agricultural and Food Chemistry. 1990 Jul;38(7):1580–5.
  83. 83. Asiedu M, Nilsen R, Lie Ø, Lied E. Effect of processing (sprouting and/or fermentation) on sorghum and maize. I: proximate composition, minerals and fatty acids. Food chemistry. 1993 Jan 1;46(4):351–3.
  84. 84. O'Deli BL, Savage JE. Effect of phytic acid on zinc availability. Proceedings of the society for experimental biology and medicine. 1960 Feb;103(2):304–6.
  85. 85. Hisayasu S, Orimo H, Migita S, Ikeda Y, Satoh K, Shinjo S, et al. Soybean protein isolate and soybean lectin inhibit iron absorption in rats. The Journal of nutrition. 1992 May 1;122(5):1190–6. pmid:1564573
  86. 86. Reddy MB, Hurrell RF, Juillerat MA, Cook JD. The influence of different protein sources on phytate inhibition of nonheme-iron absorption in humans. The American journal of clinical nutrition. 1996 Feb 1;63(2):203–7. pmid:8561061
  87. 87. Perlas LA, Gibson RS. Use of soaking to enhance the bioavailability of iron and zinc from rice‐based complementary foods used in the Philippines. Journal of the Science of Food and Agriculture. 2002 Aug 1;82(10):1115–21.
  88. 88. Sandberg AS. Bioavailability of minerals in legumes. British Journal of Nutrition. 2002 Dec;88(S3):281–5.
  89. 89. Lestienne I, Caporiccio B, Besançon P, Rochette I, Trèche S. Relative contribution of phytates, fibers, and tannins to low iron and zinc in vitro solubility in pearl millet (Pennisetum glaucum) flour and grain fractions. Journal of agricultural and food chemistry. 2005 Oct 19;53(21):8342–8. pmid:16218686
  90. 90. Greiner R, Konietzny U. Improving enzymatic reduction of myo‐inositol phosphates with inhibitory effects on mineral absorption in black beans (Phaseolus vulgaris var. preto). Journal of food processing and preservation. 1999 Sep 1;23(3):249–61.
  91. 91. Lönnerdal BO. Dietary factors influencing zinc absorption. The Journal of nutrition. 2000 May 1;130(5):1378S–83S. pmid:10801947