Skip to main content

Yeast Choice Won't Close the Aroma Gap: Budner et al. (2024), Reviewed

A controlled comparison brewed the same recipe with four different ale yeasts on sorghum extract and on barley extract. The sorghum beers ran far lower on every one of nine measured aroma compounds — and changing the yeast strain didn't significantly change that on either grain. The aroma character of a sorghum beer is set by the wort you give the yeast, not by which yeast you pitch. This strengthens existing GFB guidance; it changes none of it.

The review question​

When a sorghum beer comes out aromatically thin next to a barley beer, can yeast selection fix it? Brewers reach for strain swaps first because they are the cheapest experiment. This paper is the most direct published test of that instinct we have found.

The current GFB position​

Aroma in sorghum beer is engineered upstream: hop and yeast choices matter for fermentation character, but the site's standing doctrine is that yeast nutrition is a designed input — sorghum worts run low on FAN (the Hartwick production malts measured 46–87 mg/L against a barley working range near 140–180), and ester production is nitrogen-hungry. This review asks whether the paper confirms, conflicts with, or extends that position.

The study​

Budner, D., Carr, J., Serafini, B., Tucker, S., Dieckman-Meyer, E., Bell, L., Thompson-Witrick, K.A. "Targeted Study of the Effect of Yeast Strain on Volatile Compounds Produced in Sorghum Beer." Foods 13(22):3626, 2024. doi:10.3390/foods13223626 — open access, CC BY 4.0 (full text). Coastal Carolina University and the University of Florida.

What they did:

  • Two worts: sorghum malt extract syrup versus amber barley malt extract syrup (both Maillard Malts), same weight into ~23 L boils, fermented at 19 L. Three replicate brews per grain.
  • Four yeasts: Wyeast 1056 (American Ale), 1098 (British Ale), 1010 (American Wheat), 1214 (Belgian Abbey), fermented 14 days at room temperature, sampled at days 0, 3, 7, and 14.
  • Nine target volatiles by SPME-GC-MS: seven esters (ethyl butyrate, butyl acetate, isoamyl acetate, ethyl caproate, hexyl acetate, ethyl octanoate, ethyl decanoate), one fusel alcohol (1-octanol), and nonanal.

What it found​

  • At day 0 the worts were statistically indistinguishable on the nine compounds — the gap is made during fermentation, not inherited from the extract.
  • From day 3 onward, barley beers ran significantly higher than sorghum beers on the measured compounds (p < 0.001), and stayed higher through day 14. The sorghum fermentations produced these secondary metabolites at a much lower rate.
  • Yeast strain had no statistically significant effect on these nine compounds — in either grain. Four strains with very different reputations produced the same targeted-volatile picture.

Where it agrees with our record​

The authors' own explanation points at nitrogen: sorghum wort carries less free amino nitrogen, and low FAN suppresses yeast growth and secondary metabolism — esters above all. That is the mechanism this site's yeast-nutrition doctrine is built on, arrived at from production experience and malt analysis rather than from fermentation volatiles. Two independent routes to the same wall.

Material limitations​

Read before quoting this paper at anyone:

  • The barley control was an amber extract. Kilned differently than the sorghum extract — so the grain comparison carries a color/Maillard confound on top of the species difference. Direction of the volatile gap is unaffected (these are fermentation products, made after day 0, from worts that started indistinguishable on the measured compounds), but "barley versus sorghum" here is really "one amber barley extract versus one sorghum extract."
  • FAN was not measured. The nitrogen explanation is inference from the literature, not a measured variable in this study — the mechanism is plausible and consistent with our data, but this paper does not prove it.
  • Extract brewing, not mashing. Nothing here speaks to whole-grain sorghum mashes, enzyme choices, or process temperature effects.
  • Nine compounds is not an aroma. Targeted GC-MS sees what it aims at; there was no untargeted profile and no sensory panel. Notably, the Belgian strain's signature phenolics were not on the target list — so "no strain effect" means on these nine compounds, not "all yeasts taste the same."
  • Pitch condition unmeasured. Cell counts and viability were not determined before pitching.

Practical interpretation​

  • Don't shop for a yeast to fix thin sorghum aroma. Four strains couldn't move nine core aroma volatiles. The cheap experiment is cheap because it doesn't work.
  • Spend the effort where the mechanism lives: nutrition dosing to bring the wort's nitrogen up to what yeast secondary metabolism needs, and aroma built deliberately from hops and recipe design.
  • Yeast choice still matters for what this study didn't measure — attenuation, phenolic character, fermentation behavior. Choose strains for those reasons, not for ester rescue.

What this evidence does not establish​

That the gap is fixed. The obvious next experiment — the one this paper sets up but does not run — is the same comparison with the sorghum wort's FAN raised to barley levels by nutrient dosing. If the ester gap closes, the nitrogen mechanism is confirmed and the practical fix is validated; if it doesn't, something more interesting is wrong. That experiment is squarely within homebrew-scale reach, and it is exactly the shape of work our Research Asks exist to place.

Implications for GFB guidance​

No page changes. The hops-and-yeast and yeast-nutrition pages already say what this paper supports; this review becomes the citable external evidence behind them.

Source and Validation Notes​

Reviewed from the open-access full text (CC BY 4.0) linked above. Characterizations of the study's design and findings are ours; specific concentration values are deliberately not reproduced here — the direction and significance of the differences are the load-bearing findings, and readers who need the numbers should take them from the paper itself. If the authors find this characterization wrong anywhere, that is exactly the correction we want.