Vegan bakery: Researchers test wheat bran albumin as egg replacer in cookies and cakes, with comparable spread, volume and density
Water-soluble albumin extracted from wheat bran, a milling by-product, has been used as a complete egg replacement in cookies and cakes. Researchers in the Netherlands report comparable spread factor, cake volume and density, and higher antioxidant activity in cookies. They also flag colour shifts and marked texture differences.
Albumin recovered from wheat bran can replace eggs entirely in cookies and cakes, with a comparable spread factor, cake volume and density, according to a study published in Applied Food Research. The open access paper comes from the University of Groningen and Royal Koopmans in Leeuwarden, the Netherlands. It compares egg-based formulations with ones made using water-soluble albumin extracted from ultra-fine wheat bran, a by-product of flour milling.
The authors, Varshini Krishnamoorthy, Jan Krijgsheld and Gert-Jan W Euverink, extracted the albumin at pilot scale. In summary they report: “These findings underscore the use of wheat bran albumin as an egg replacer in baked goods, aligning with consumer demand and sustainability goals.”
Implications for manufacturers
The paper’s commercial case rests on the extraction process. The authors describe a water-based method that uses no organic solvents, and they say it is inexpensive to run. In their words: “The practical applicability of the water-based albumin extraction technology is high, as it is cost-effective and avoids the use of organic solvents.”
They also address likely engineering problems. Column clogging from dust or fine starch, they write, can be mitigated with a moving-bed extractor, and “small-scale albumin extraction poses no major challenges.” Wheat bran is a low-cost raw material containing 13-18% protein, according to the authors, who note that it is often undervalued or treated as waste. For millers, the work offers a route to valorise a side stream. For bakery and confectionery manufacturers, it offers a plant-based ingredient for vegan lines. They note that although eggs perform several structural roles in cookies and cakes, including aeration, foaming and emulsion stability, they have been associated with health concerns including high cholesterol.
Regarding the scope of the study, the authors note: “While safety and allergenicity assessments were beyond the scope of this study, the technology provides a cheaper alternative to eggs, making the replacement of eggs with albumin in vegan cookies and cakes feasible for commercialisation.” No sensory evaluation was carried out, and the albumin products differed from the egg controls in colour and texture, as detailed below.
Physical appearance of cakes formulated with (a) egg and (b) albumin
© Krishnamoorthy, V et. al doi: j.afres.2026.102400
Extraction and formulation
The team used Osborne fractionation, which separates proteins by solubility, to isolate the water-soluble albumin. Ultra-fine wheat bran supplied by Koopmans Meel B.V. was combined with water at a 1:12 ratio in a 1000 L stainless steel reactor and stirred overnight at room temperature. The mixture was then concentrated through a spiral nanofiltration membrane with a 150-300 Da cut-off, operated at a feed pressure of 5-6 bar. After centrifugation at 9,000 x g for 35 minutes to remove bran and starch, the protein solution was concentrated further with a 10 kDa ultrafiltration membrane and freeze-dried. Protein content was determined by Bradford assay. The extraction yield was 92%, giving 1.15 kg of albumin from 40 kg of ultra-fine wheat bran.
Egg was replaced completely in both products. The albumin cookies used 9 g of albumin against 26 g of egg in the control. The albumin cake used 11 g against 31.25 g of egg. Oil and vegan milk were blended with the albumin to bring fat and water content close to that of egg.
The cookie formulations also differed in the fat used. The control contained dairy butter at 82% fat, while the albumin cookies contained vegan butter at 72% fat. Both products were baked at 175 °C, the cookies for 9-11 minutes and the cakes for 15-20 minutes. Measurements were made in triplicate, with Student’s t-test for pairwise comparisons and two-way ANOVA for the antioxidant assay.
Cookie performance
Spread factor, a key indicator of dough behaviour and cookie geometry, was not significantly affected by the replacement. Egg cookies measured 4.6 ± 0.36 and albumin cookies 4.7 ± 0.03, with no significant difference in width or thickness. The authors interpret this as indicating a low-viscosity dough that yields desirable cookies.
Soluble protein in the baked cookies was similar between formulations, at 3.0 ± 0.12 mg/g for egg and 3.5 ± 0.35 mg/g for albumin (p ≥ 0.05). Protein in the doughs did differ, at 7.9 ± 0.54 mg/g for egg and 6.4 ± 0.07 mg/g for albumin. The authors attribute the fall in soluble protein after baking to Maillard reactions, denaturation and aggregation, lipid-protein complex formation, moisture loss and breakdown of labile protein.
The pH divergence was marked. Egg cookies rose from 7.9 in the dough to 8.9 after baking, which the authors link to the alkalinity of egg white and to sodium carbonate formed from the baking soda. Albumin cookies stayed mildly acidic at 6.7. Baked moisture content was significantly lower in the albumin cookies, at 5.9 ± 0.74% against 9.0 ± 1.03%. The authors caution that oven-drying at 105 °C for 3 hours measures total removable moisture and cannot be used to infer differences in water-binding capacity.
Texture showed the largest difference. Hardness was 42.2 ± 6.00 N for egg cookies and 285.2 ± 1.51 N for albumin cookies (p < 0.001). The authors write that “albumin cookies were markedly harder, reflecting a rigid or denser structure.” They suggest higher protein content and lower fat content as possible contributors, the latter because reduced fat can increase sugar crystallisation and so reduce tenderness. Fracturability did not differ significantly, at 5.0 ± 0.00 mm for egg and 4.7 ± 0.35 mm for albumin, indicating comparable brittleness.
[/av_textblock]
[av_hr class='default' icon_select='yes' icon='star' position='center' shadow='no-shadow' height='50' custom_border='av-border-thin' custom_width='50px' custom_margin_top='30px' custom_margin_bottom='30px' custom_border_color='' custom_icon_color='' av-desktop-hide='' av-medium-hide='' av-small-hide='' av-mini-hide='' id='' custom_class='' template_class='' av_uid='av-1lmro16' sc_version='1.0' font='svg_entypo-fontello']
[av_image src='https://fei-online.com/wp-content/uploads/sites/4/2026/10/cookies_comparisson-1030x239.jpg' attachment='14444' attachment_size='large' src_dynamic='' copyright='' caption='' image_size='' styling='' box_shadow='none' box_shadow_width='10' box_shadow_color='' align='center' font_size='' overlay_opacity='0.4' overlay_color='#000000' overlay_text_color='#ffffff' link='' link_dynamic='' target='' animation='no-animation' animation_duration='' animation_custom_bg_color='' animation_z_index_curtain='100' parallax_parallax='' parallax_parallax_speed='' av-desktop-parallax_parallax='' av-desktop-parallax_parallax_speed='' av-medium-parallax_parallax='' av-medium-parallax_parallax_speed='' av-small-parallax_parallax='' av-small-parallax_parallax_speed='' av-mini-parallax_parallax='' av-mini-parallax_parallax_speed='' hover='' blur_image='' grayscale_image='' fade_image='' appearance='' css_position='' css_position_location=',,,' css_position_z_index='' av-desktop-css_position='' av-desktop-css_position_location=',,,' av-desktop-css_position_z_index='' av-medium-css_position='' av-medium-css_position_location=',,,' av-medium-css_position_z_index='' av-small-css_position='' av-small-css_position_location=',,,' av-small-css_position_z_index='' av-mini-css_position='' av-mini-css_position_location=',,,' av-mini-css_position_z_index='' transform_perspective='' transform_rotation=',,,' transform_scale=',,' transform_skew=',' transform_translate=',,' av-desktop-transform_perspective='' av-desktop-transform_rotation=',,,' av-desktop-transform_scale=',,' av-desktop-transform_skew=',' av-desktop-transform_translate=',,' av-medium-transform_perspective='' av-medium-transform_rotation=',,,' av-medium-transform_scale=',,' av-medium-transform_skew=',' av-medium-transform_translate=',,' av-small-transform_perspective='' av-small-transform_rotation=',,,' av-small-transform_scale=',,' av-small-transform_skew=',' av-small-transform_translate=',,' av-mini-transform_perspective='' av-mini-transform_rotation=',,,' av-mini-transform_scale=',,' av-mini-transform_skew=',' av-mini-transform_translate=',,' mask_overlay='' mask_overlay_shape='blob' mask_overlay_size='contain' mask_overlay_scale='100%' mask_overlay_position='center center' mask_overlay_repeat='no-repeat' mask_overlay_rotate='' mask_overlay_rad_shape='circle' mask_overlay_rad_position='center center' mask_overlay_opacity1='0' mask_overlay_opacity2='1' mask_overlay_opacity3='' title_attr='' alt_attr='' img_scrset='' lazy_loading='disabled' id='' custom_class='' template_class='' av_element_hidden_in_editor='0' av_uid='av-mvjec8l3' sc_version='1.0' admin_preview_bg=''][/av_image]
[av_textblock fold_type='' fold_height='' fold_more='Read more' fold_less='Read less' fold_text_style='' fold_btn_align='' textblock_styling_align='' textblock_styling='' textblock_styling_gap='' textblock_styling_mobile='' size='' av-desktop-font-size='' av-medium-font-size='' av-small-font-size='' av-mini-font-size='' font_color='' color='' background='bg_color' background_color='' background_gradient_direction='vertical' background_gradient_color1='#000000' background_gradient_color2='#ffffff' background_gradient_color3='' padding='' padding_sync='true' av-desktop-padding='' av-desktop-padding_sync='true' av-medium-padding='' av-medium-padding_sync='true' av-small-padding='' av-small-padding_sync='true' av-mini-padding='' av-mini-padding_sync='true' fold_overlay_color='' fold_text_color='' fold_btn_color='theme-color' fold_btn_bg_color='' fold_btn_font_color='' size-btn-text='' av-desktop-font-size-btn-text='' av-medium-font-size-btn-text='' av-small-font-size-btn-text='' av-mini-font-size-btn-text='' reveal='' reveal_speed='medium' container_tag='' fold_timer='' z_index_fold='' id='' custom_class='' template_class='' av_uid='av-mvjecoj5' sc_version='1.0' admin_preview_bg='']
Physical appearance of cookies formulated with (a) egg and (b) albumin
© Krishnamoorthy, V et. al doi: j.afres.2026.102400
Cake structure and texture
For cakes, the albumin formulation gave comparable physical attributes to the egg control. The researchers found no significant differences in weight (30.6 ± 2.01 g for egg and 27.6 ± 1.63 g for albumin), height (1.4 against 1.8 cm), volume (40.7 against 44.7 cm³), specific volume (1.3 against 1.6 cm³/g) or density (0.7 against 0.6 g/cm³). The albumin cake’s volume and height values showed wide variation, with standard deviations of 10.99 cm³ and 0.43 cm respectively.
Baked protein content was significantly higher in the albumin cake, at 4.1 ± 0.36 mg/g against 3.2 ± 0.25 mg/g for egg, even though the egg batter was higher in soluble protein (9.3 against 5.5 mg/g). Baked moisture was similar, at 15.3 ± 1.66% for egg and 14.5 ± 2.91% for albumin.
The authors attribute the volume and density results to the functional properties of the albumin: “Thus, albumin acts as an emulsifier and effectively stabilises the entrapped air bubbles, contributing to a favourable cake density.” They also note that cakes made with other plant-based replacers, including tofu and chia seeds, have shown reduced volume in earlier studies.
Baking loss was higher in the albumin cake, at 12.4 ± 2.88% against 7.2 ± 1.86%. The authors link this to a possibly lower water-binding capacity and cite its relevance to weight-related regulatory compliance and production economics. For manufacturers working to declared weights, this is a parameter to monitor in any scale-up.
Texture profile analysis showed a much softer crumb in the albumin cake. Hardness was 4.3 ± 1.00 N against 43.2 ± 12.08 N for egg, and chewiness was 2.2 ± 0.55 N against 28.5 ± 9.63 N. Springiness was marginally higher in the albumin cake (0.9 against 0.8), and cohesiveness slightly lower (0.5 against 0.6). The authors point out that the albumin cake “was softer and required considerably less energy to masticate,” and they describe its texture profile as more desirable. No sensory panel was used to test that judgement.
Colour and antioxidant activity
Colour is the most visible trade-off. The authors measured surface colour in the CIELAB system, where L* denotes lightness (0 is black, 100 is white), a* the green-red axis and b* the blue-yellow axis. Albumin products were darker and redder than the egg controls. Cookie lightness (L*) fell from 69.4 ± 3.84 to 59.7 ± 4.99, and redness (a*) rose from 4.0 to 8.4. In cakes, L* fell from 68.9 ± 3.88 to 49.8 ± 0.45, and a* rose from 2.6 to 11.7. The yellowness index was also higher with albumin, at 71.0 against 58.7 in cookies and 88.7 against 50.0 in cakes.
The total colour difference (ΔE), which combines the differences in L*, a* and b*, was 11.0 ± 4.7 for albumin cookies and 22.6 ± 1.1 for albumin cakes. The authors describe both as visually perceptible. The browning index and the blue-yellow value (b*) did not differ significantly between formulations, so the colour shift was driven mainly by lightness and redness. The authors suggest the higher protein concentration may have promoted Maillard browning by increasing the availability of reactive amino groups. They also note that the characteristic yellow colour of baked goods originates from carotene and xanthophylls in egg yolk.
Antioxidant capacity was measured by ABTS radical scavenging at concentrations of 0.125-3 mg/mL, with ascorbic acid as a control. Albumin cookies showed inhibition of 29.6-87.7% across that range, rising with concentration, and differed significantly from egg cookies (p < 0.001). Albumin cakes ranged from 17.3 to 82.2%, but the difference from egg cakes was not statistically significant (p > 0.05).
The authors attribute the higher activity to Maillard reaction products formed during baking, which can act as hydrogen donors. They also suggest that free polyphenols in wheat bran may have been co-extracted with the albumin, and that interactions between albumin and phenolic compounds may have improved antioxidant stability and extractability.
Limitations and next steps
The authors acknowledge several constraints. Safety and allergenicity were not assessed, and no sensory evaluation was performed. The albumin formulations also contained additional oil and soy milk, so the comparison with egg is not like-for-like in ingredient terms. The cookie formulations additionally differed in butter type.
The authors add that “the absence of egg-derived phospholipids could affect crumb structure and sensory characteristics, despite similar overall lipid content.” On consumer acceptance, they write: “Sensory evaluation was not performed in this study; given the relevance of sensory characteristics for consumer acceptance, future studies should address this aspect.”
Conclusion
For formulators, the data indicate where reformulation effort would be needed: cookie hardness, cake baking loss and the colour shift. The authors conclude that “this study promotes the sustainable use of underutilised milling by-products in functional food formulations as egg replacers to meet the growing demand of the vegan population.”
Source
Krishnamoorthy, V., Krijgsheld, J., & Euverink, G.-J. W. (2026). Towards vegan bakery products: Plant protein as an egg replacer in cookies and cakes. Applied Food Research, 6, 102400.
https://doi.org/10.1016/j.afres.2026.102400