Pushing past yeast convention could boost craft beer flavour and adoption
In this instalment of The Beerologist, I will describe a scientific paper by Prof. Dr Diego Bonatto that I recently found on bioRxiv (https://www.biorxiv.org/content/10.1101/2020.07.17.209171v2). The manuscript has been peer-reviewed and is now published here.
The average price of a six-pack of (US) craft beer is $12. For only $5 per month (about 3 bottles of craft beer), you support our mission: To keep you up to date with the latest brew research. Why not sign up and give us a go?
What is this paper about?
The ecology of yeast in brewing. As it happens, the principles of ecology can be applied to many systems and problems. Brewing is not any different. But before delving into this work, let’s talk about what ecology is and how it applies to brewing beer.
Ecology is a discipline in biology that studies the interactions between organisms and the physical environment. Ecologists are interested in how (micro) organisms co-exist in a given system and how external physical factors affect the ecosystem and its inhabitants (and vice versa). In this context, the craft beer industry can be considered a large ecosystem, composed of a set of habitats (beer styles or categories), each of which will contain a number of related recipes that each call for a particular yeast strain. A critical question thus becomes: (To what extent) does beer style dictate the choice of yeast strain? While in some cases yeast strain selection is obvious (for example, using a lager yeast strain for making a lager and an ale strain for an ale) it is not clear what drives yeast strain selection in particular beer categories. By treating this puzzle as an ecological question, Diego Bonatto was able to attack this problem for the first time using a large database of 121,528 recipes from the brewers’ friend website (https://www.brewersfriend.com).
Beer type and yeast formulation matters
Using the recipe dataset, the first question that needed to be asked was: What is the yeast “population structure” across the entire recipe ecosystem and what attributes jump out? Counting strains in ale and lager beer recipes revealed that there were many more different ale yeasts than lager yeasts (and recipes). In addition, when looking at specific categories, it appeared that some beer styles featured many more different yeast strains than others. Furthermore, it became immediately apparent that in craft brewing, dried yeast formulations are used much more often than their liquid counterparts. Interestingly, however, the majority of yeast strains come in liquid formulations, an observation that the author thinks is due to the fact that some strain types are more tolerant of the drying process than others.
A large and largely complete dataset allows for the identification of factors (ABV, IBU, Final Gravity, etc.) that are associated with yeast strain. Identifying correlation between strains and beer attributes is powerful since it can help you tease out the biological attributes that help a given strain perform well (e.g. efficiently convert sugar) and/or produce the flavour profiles that are associated with a given beer style. These analyses revealed that only IBU (bitterness imparted by hop additions) was correlated with a high number of yeast strains and recipes.
Perhaps the most relevant experiment in this work was the assessment of yeast richness and evenness amongst beer categories. In ecological terms, yeast richness refers to the number of different yeast strains in a given beer category. Yeast evenness describes the level of strain distribution across all beer styles and their abundance in each category. A category can be very diverse (many different strains that have similar levels of abundance in a category) or feature low diversity (a category is dominated by only a few strains). By calculating strain richness and evenness in each category and for each strain, we are able to assess the level of dominance or in other words, the likelihood with which a given strain is associated with a beer style.
What does this work reveal?
In a nutshell and to summarise, craft brewers tend to select dry yeast formulation despite the availability of a wide selection of liquid yeast strains. Interestingly, in both lager (five out of nine) and ale beer categories (ten out of 25), a significant number of beer styles featured low levels of yeast diversity (see table below). These results suggest that either the beer ingredients and brew conditions are not conducive to performance for many strains. Alternatively, these results indicate that for reasons unknown, the use of alternative yeast strains has not been explored for some styles.
These analyses allow a targeted approach to introduce new strains and flavour profiles into existing beer styles. Testing yeast strains that are not typically found in current recipes, but perform well in established formulations will increase the diversity and range of beers within a given style.
Good answers raise more questions
With any interesting piece of research, new insights lead to new questions, some of which are described below.
What are the origins of the recipes in the database and does it have an impact on the results? The availability of this unique dataset (121,528 recipes) offers opportunities but still has some limitations. The sheer size of the dataset allows robust statistical analyses, revealing relationships that are supported with high levels of certainty. The problem, however, is that we do not know to what extent this database represent all recipes out there. While the database is based in the USA, we do not know where all the recipes come from. In addition, with two of the major yeast suppliers based in the USA and Europe respectively, it is unclear whether there is any geographical factor at play. Similarly, is there a historical factor at play that could affect the wider conclusions? Without the appropriate data, these questions are hard to answer. Importantly, all conclusions are supported by experimental evidence. Finally, do craft brewers follow these recipes to the letter or do they replace yeast strains or their formulations at their own accord? Surveys targeting active brewers, using these recipes, will help answer the last question.
Finally, there is one other aspect that warrants some thought. While we now can assess diversity from an ecological perspective, how does this translate to genetic diversity? On the molecular and genetic level, there are relationships between strains that are yet to be uncovered. If the ecology of a category says that there is high levels of diversity, what does that mean on the genetic level? Are all strains distinct from each other or do they share a particular ancestor? Genetic and/or genome wide analyses, comparing the genetic makeup of a set of strains, will help answer this important question. The work described here, certainly provides a solid foundation from which further detailed studies can be designed.
In my next post, I will share with you a Q&A exchange between Diego Bonatto and myself about the paper and his reasons for embarking on this research. In addition, we will write about or how the data and results from this work can be used to improve beer diversity and quality.
If you like this post and if you do not receive my free emails, please consider joining my subscriber list and hear about the latest brewing-associate science.
Edgar, the Beerologist