Meeting Flour Specifications

In my previous article from the Third Quarter issue of Milling Journal, “Continued Observations on Drought and Wheat Performance,” the impact of drought on wheat performance in Kansas was reviewed.

It was concluded that wheat sourcing for flour production to meet bakery specifications is a quite challenging. Flour specifications and baking properties are influenced by both protein quantity and quality, in addition to the presence of other important flour components, including starch, damaged starch, and pentosans (cell wall material).

Bakery flour specifications often include such factors as mixing time, tolerance, dough handling properties, and water absorption. Additionally, baking properties affected include oxidation requirements, loaf volume, and crumb characteristics of bread.

This brief report focuses on protein quantity and quality impact on loaf volume and absorption for bread flour. As discussed previously, the wheat crop harvest is composed of many wheat varieties from year to year and across the wheat production region. In addition, environmental and agronomic factors vary across wheat production regions. These factors influence both protein quality and quality.

Flour protein quantity is an attribute measured based on the presence of nitrogen which is part of the amino acids making up protein. The proteins most important to wheat flour and bread baking are those forming gluten. Not all proteins present in wheat or flour are gluten-forming proteins.

There are several test methods available to quantify the presence of protein and especially gluten-forming protein in wheat and flour. These tests are valuable in that they offer a more rapid prediction of protein quality than baking bread. However, the final arbiter of protein quality is making bread. The Finney and Barmore reference provided is often identified as the early stage of protein quantity and quality assessment for wheat varieties.

The Finny and Barmore study identified wheat variety performance for a collection of hard winter and hard spring wheat flour varieties by measuring loaf volume as protein content increased. Data from the study is often cited to demonstrate the difference in protein quantity and quality. Table 1 on p. 6 shows the protein quality assessment and slope estimated for loaf volume at 12% and 18% protein. Greater slope represents improved protein quality when compared to wheat or wheat varieties with a lesser slope. Using the slope and endpoints from Table 1, Figure 1 on p. 7 was developed showing points between 12% and 18% at 0.5% increments. Such a linear graph may lead a miller to think that increasing protein content results in increase flour attributes whether it is absorption or loaf volume.

Commercial wheat is a composite of many wheat varieties harvested around any given collection point at local elevators. Wheat from local elevators may be further co-mingled at a terminal location before delivery to the mill. Even if wheat is delivered direct from the producer, sources used to create your mill mix result in a mix of many varieties grown under many different conditions. 

Figure 2 on p. 7 was created to show loaf volume estimates for a crop year given a blend of the wheat characterizations identified in Table 1 at each of the various protein levels between 12% and 18% in 0.5% increments. The best fit linear regression equation and R2 value are shown in the figure.

An R2 value of 0.659 suggests the linear regression equation explains approximately 65.9% of the variation in loaf volume. Remember, R2 is the proportion of variation in the dependent variable (loaf volume) that can be attributed to the independent variable (flour protein).

Clearly, there are factors at play other than protein quantity that the regression equation does not explain. A miller looking at the graph might assume increasing protein content of the wheat mix will always result in increased loaf volume. There is a reasonable chance that increasing wheat mix protein will increase loaf volume but may in fact result in decreased loaf volume depending on wheat protein quality. The good news is that increased loaf volume can reasonably be expected with blends of known varieties and growing conditions even across production years.
Wheat variety protein quality relationships are a consistent trait across various growing conditions and production years. A wheat variety represents genetic potential whose expression is controlled by environmental conditions and agronomic practice.

Analyzing the 2023 HRW Wheat Crop
The study of Finney and Barmore is important, but how does it relate to today’s hard red winter (HRW) wheat crop? Figure 3 on p. 8 shows the relationship between flour protein and loaf volume for the 2023 HRW wheat crop. The best fit linear regression equation and R2 value are shown in Figure 3. An R2 value of 0.6775 suggests the linear regression equation explains approximately 67.8% of the variation in loaf volume as a function of protein level. The data collected and relationship between flour protein and loaf volume in 2023 is like that obtained some 75 years ago in 1948. So, the point is, to assist your bakery customer to increase loaf volume, it will require more than simply increasing protein content in the flour!

It is unlikely your customer’s flour specification holds you to a specified loaf volume standard. Often, water absorption, mix times, and tolerance are specified flour attributes that may be tested in less time than a baking performance test.

Unlike baking performance tests (a direct finished product performance test), these tests are more predictive in nature, utilizing a narrower set of flour attributes, rather than the matrix of attributes and processing factors considered in baking performance measures. They are, however, the standard by which flour specifications are measured and considered critical to successful bakery flour performance.  

Figure 4 on p. 9 shows the relationship between flour protein content and Farinograph water absorption for HRW composites harvested in 2023. The best fit linear regression line is shown in red with the linear equation and R2 value are shown in Figure 4. An R2 value of 0.4121 suggests the linear regression equation explains approximately 41.2% of the variation in Farinograph water absorption as a function of protein level. With respect to absorption, consider that protein absorbs approximately 1.8 times its weight in water, damaged starch absorbs four times its weight in water (compared to native starch which absorbs only 0.4 times its weight in water), and pentosans absorb approximately 10 times their weight in water.
Absorption characteristics of flour components impact flour performance and can be measured with laboratory tests such as Solvent Retention Capacity (SRC).

The Park et al reference in evaluating the impact of various protein fractions reported a correlation value of r=0.45 for the relationship between flour protein content and absorption. The correlation value is represented by r and a number between -1 and 1 identifying the relationship between two variables, in this case flour protein (independent variable) and absorption (dependent variable). A value of zero (0) indicates there is no relationship while a value of either 1 or -1 indicates a near perfect relationship.

The sign of the value indicates the direction of the relationship. A positive (+) relationship indicates as the independent variable increases the dependent variable also increases, with a negative (-) value as the independent variable increases the dependent variable decreases.
The r value reported results in a R2 value of 0.2025, indicating the regression equation for water absorption as a function of flour protein explains 20.25% of the absorption variation observed. With respect to loaf volume as a function of flour protein content Park et al reported r=0.8, resulting in an R2 value of 0.64, explaining 64% of the variation in the linear regression.

The limited number of varieties evaluated may have impacted the estimated relationship observed by Park et al while the crop report data shown in Figures 3 and 4 utilized 70 composite samples each with many undisclosed varieties grown in different environments under varying agronomic conditions. Perhaps Figures 3 and 4 represent the reality of the commercial wheat quality in the United States for loaf volume and water absorption relationships to wheat flour protein.

It is important to consider, however, that each composite was collected from a small and reasonably well-defined draw area or grain shed as can be identified on the Plains Grains Inc. (PGI) website (plainsgrains.org/wheat-data-maps). For detailed information about the crop quality in individual grain sheds, contact PGI at plainsgrain@gmail.com to request access.
As the number of flour mills in the United States has decreased, milling capacity for each site has increased. The increase in capacity has increased the local wheat draw area for a given origin mill site. Many milling sites have been modified from milling a specific class of wheat to alternating or swinging between classes on a single milling unit.

Alternatively, specific milling units on-site are dedicated to milling a wheat class produced well away from wheat origin. Some refer to these mills as “destination mills,” as they are built closer to population centers where the wheat is destined to be milled and utilized. Some suggested the development of destination mills resulted when “milling in transit” was eliminated, increasing the cost of transporting value-added flour to customers near population centers.
Additionally, at any time during the year, 26.5-35.4% of the wheat crop is held in on-farm storage to be released at the producer’s discretion. Wheat selection has never been more challenging and requires careful evaluation.

Conclusion
The goal of this brief report was to focus on protein quantity and quality impact on loaf volume and absorption for bread flour from a commercial perspective. That a linear model of flour protein and loaf volume can be developed with a positive correlation does not guarantee an increase loaf volume due to increased flour protein level. Protein quantity and quality must be considered in the wheat selection process.
Likewise, that a linear model of flour protein and loaf volume can be developed with a positive correlation does not guarantee increased absorption due to increased protein level. The values of these linear models should not be discarded as they do provide directional assessment of protein impact on both loaf volume and absorption. Protein quality is as important if not more so than as protein quantity.

When making a cash wheat purchase, study further back in the supply chain to wheat origin to assess wheat properties, including flour and dough properties and functionality may benefit the milling operation and ability to meet specifications.

The matrix of flour quality measurements and attributes is complex, and the impact of various components are yet to be identified and interactions evaluated. The ability to meet customer specifications, however, begins with wheat selection. More must be learned to manage wheat selection and optimization. We must have a better understanding of the size and distribution of compound particles in each flour stream, as well as individual starch granule size and distribution along with variations in cell wall thickness, as it may affect the amount of cell wall present hence pentosans level in flour.

Along with starch damage comes the disruption of cell wall material – unknown is the amount of damage to the cell wall and liberation of pentosans. With better understanding, grinding techniques may be developed to better manage disruption of starch granule, cell wall material, and proportion of compound particles in flour.

With better understanding, we can improve on the long-standing solution of increasing protein levels and grinding harder to increase flour absorption or loaf volume.

Perhaps artificial intelligence can be used to analyzed wheat and flour data to better understand our options. After all, flour quality is more than a two-dimensional comparison between two variables or three-dimensional comparison in a response surface for three variables or four perhaps four-variables considered shown on a flat piece of paper as used in SRC analysis. Perhaps flour quality can be best assessed by considering each face of a multi-faced die as an individual factor and determining where in the interior successful flour properties reside.

Dr. Jeff Gwirtz is CEO of JAG Services, Inc., an international consulting company in Lawrence, KS; 785-341-2371; jeff@jagsi.com. He also is adjunct professor in the Department of Grain Science and Industry at Kansas State University, Manhattan.