Category Archives: Process

Beer Yeast for Celiacs, Part I

One of my big beer-related interests is brewing with “alternative” grains, partly because of curiosity but also because my wife has celiac disease and likes beer. For the unfamiliar, the treatment for celiac disease is to avoid all gluten-containing foods, whether they are made with gluten-containing ingredients (barley, wheat and rye and their derivatives) or potentially cross-contaminated with remnants of gluten-containing ingredients. This can be a lot more complicated than it seems (partly due to not being taken seriously, thanks to fad dieters and health food nuts co-opting your involuntary, medically-necessary diet, but I digress…) and the consequences of not doing so are serious, far, far beyond than the stomach discomfort most would associate with it.

Conventional beer is out of the question as it is made with at least one of the three problematic grains. Filtering and/or enzymatically breaking down the leftover gluten in beer is possible — beers like Estrella Daura or the Omission line are made this way and are tested for safety. Levels below 20ppm are held to be safe for celiacs, and done well these treatments are effective. But this is at the very least unreliable at the homebrew level since we can’t test the gluten content in the finished product short of sending samples to a lab.

Fortunately, we can “simply” malt and brew with other grains and pseudocereals, like millet, rice, sorghum, maize, buckwheat and quinoa, all of which are naturally gluten-free. Since hops are naturally gluten-free, wort production is taken care of–although often with much more difficulty than barley beer and with a distinct flavor from each of the different grains.

That brings us to yeast. Commercial liquid yeast manufacturers use barley wort as the base in their growth medium, and the finished product ends up containing some. The levels are low, low enough that we can ignore it: about 120ppm for Wyeast, and an extremely low 12ppm for White Labs. When diluted in several gallons of wort, that gluten content is surely under the aforementioned accepted maximum of 20ppm. That is great, for sure. Some celiacs are fine with that level of assurance. But we are not perfect (or perfectly rational) beings, and while we may know it is not going to be a problem, sometimes whatever doubt lingers in our minds is enough to ruin the experience of having a nice beer.

To fully eliminate any apprehension, we can use dry yeasts from (at least) Danstar, Safale and Mangrove Jack’s that don’t use barley wort as their medium. At least for Danstar and bread yeast manufacturers, molasses is the sugar source in growth medium. But that is further restricting due to the limited number of strains available. Not that they aren’t good, of course, but there are just so many good liquid ones that it is too important to ignore.

The good news is that yeast ranching with gluten-free media is trivial. In fact, I have mostly used only such media from day one and use them exclusively on all slants. This opens up the entire world of liquid yeast for gluten-free brewers, because we can culture the yeast taking essentially nothing but the cells themselves from the original tube/pack. Plating and further culturing is then done on gluten-free media and the resulting beer is 100% free of any contamination, provided standard hygiene is practiced.

Potato-based media are completely fine and require no changes. But even malt-based media are easy to adapt. Briess white sorghum syrup, which is made from unmalted sorghum grain that is enzymatically mashed (and thus totally different from sorghum cane syrup/molasses), is a 1:1 replacement for LME. Most malt medium recipes call for DME due to its ease of use, but we can swap in LME and thus sorghum grain syrup by accounting for the 20% water content. Some might call into question the lower FAN content of the Briess syrup relative to standard LME (as indicated by Briess’s spec sheets) but I have not found that to be a problem, and end up using yeast nutrient in starters anyway. So, that’s pretty much all there is to it.

Part II will contain a few concrete recipes and go over a few of the idiosyncrasies.

Off-flavor Follow-up

In my last post on the subject, I mentioned I had just brewed a very light cream ale, a beer in which any of this off-flavor would be immediately noticeable. Sadly, I was right about that! It is quite bad. Like my last post, this is mostly a “brain dump” and thus not quite as coherent as it could be. As always, any help is appreciated.

Given the good quality of the water I brewed this with, I am fairly confident the chlorine or pre-existing chlorophenol hypothesis is out. I do have a few new data points to work with and could instead be looking at oxidation. I had previously discounted it due to the fact the flavor develops almost right away. Generally I’ve seen it stated that it takes a long time for oxidation to show up. But there is a little more this time to suggest it.

At bottling, I noticed none of the specific off-flavor, although there was a bit of acetaldehyde. Out of curiosity I opened the last bottle (only about 3/4ths filled) only two days after bottling. It is already quite carbonated, and it poured with a lot of foam. I immediately got the off-smell and tasted it right away, too. This time I thought maybe I was falsely attributing plastic notes due to reading about such an off-flavor online and my shaky confidence about the water. Although it’s not quite “wet cardboard” or other descriptions of oxidation, there’s perhaps a general staleness to it.

The other half of the 2.5 gallon batch went into a Cubitainer (polypin) with half as much priming sugar. I’ve sampled a few pours despite it being only a little carbonated, and so far there is no real hint of the specific off-flavor or smell there. But there is still that general staleness, I think. So perhaps the carbonation is enhancing my perception of the smell, and perhaps then the smell enhances my perception of the flavor.

That is consistent with past experiences, where I have tasted this during active fermentation or after bottling, but not in the wort or flat beer. Indeed, after taking the bottled beer and pouring it repeatedly between two glasses I started to notice it less and less. So although there’s no reason to think it is a result of overcarbonation or anything like that, it’s probably connected to CO2 as far as my perception is concerned.

There were many opportunities for oxidation in this batch:

  • First, every beer I’ve made from this 10 lb. bag of pale malt has been bad (#3 & #4 from the last post and this one) Although I am crushing it at home and in used it once right when I got home from the LHBS, it’s possible it was old and stale when I bought it. The same goes for the chichas, which were made with pre-crushed malted corn from who knows when, and the first problematic beer which was made with six-month-old precrushed malt. This is supposed to be quite unlikely, but definitely possible. It would be consistent with the smell & flavor being there right from the get-go. However, the malt seems to taste fine raw.
  • Second, in this instance I tried a BIAB, no-sparge mash and let a lot of it drip down from the bag while squeezing it out. Other people have done this many times and report good results, and other beers where I have splashed or dripped a lot (like my popcorn ale and millet saison) had none of this. But hot-side aeration is not impossible, just improbable as far as I can tell so it’s a possibility.
  • Third, and perhaps most damning, I intended to do a short cold crash but ended up effectively doing a month-long lagering since the yeast took forever to drop out. Since it was supposed to be short I replaced the airlock with a piece of foil and a rubber band to avoid suckback. That amount of time with that amount of headspace (half of a 5 gallon Better Bottle) can easily cause oxidation, I think.

Without further experiments I can’t quite discount the possibility that something else is wrong and causing the off-flavor/smell in addition to this beer being oxidized and stale. I can test the malt with side-by-side small mashes with the same water and new malt. I can intentionally splash the heck out of another mash to test that. I am not super keen on wasting a month intentionally oxidizing a beer, but hey, it may be necessary.

This is all quite discouraging, of course, but I intend to figure it out. Thankfully I have had good beers before and in between the bad, so I know it’s not impossible.

DCY02 & DCY03: DCambic Isolates

DCY02 and 03 Fermentation in Progress

I am happy to report that isolating the yeasts from The Mad Fermentationist’s DCambic was a success. I have written about it previously and had some issues at first. A lot of stuff going on in life then kept me from the project for a few months but eventually I got the time to get it going again. I have been posting off and on about it on Twitter, but it’s time I document it here.

The key to success the second time was growing a liquid culture first from a sample Mike gave me, rather than plating it straight from the sample. After a few step-up culture iterations I was able to isolate four different yeasts from the beer, instead of just the one I got the first plating attempt. Two were useless for brewing: one is possibly Rhodotorula and the other is another yeast that does not ferment. The others were almost certainly Brettanomyces.

I can tell them apart on a plate because one (DCY02) grows white colonies and the other (03) tan colonies. Liquid cultures are also markedly different: DCY02 forms a pellicle very quickly, smells very funky and drops out fairly quickly, whereas DCY03 takes much longer to form a thinner pellicle, smells very fruity and drops out a lot slower. Taste profiles from the starter are along the same lines as the smell.

I made two simple extract beers in April (1.040 of pilsner DME) and the picture above is of fermentation about two weeks in. They were slow to start but that may be due to low pitching rates. In any case, I have not yet bottled these as they seemed to reactivate after I moved in late May. Perhaps just shaking them a bit roused enough yeast into suspension to do something, and I want to make sure they’re done before I bottle. (I will update this once they are done.)

Strain Descriptions

DCY02: The “white” isolate, based on colony appearance. This has a funkier Brett character with less of the fruity qualities. It produces a pellicle very quickly, and during fermentation produces a thick foam cap that looks more like Sherry flor than ale yeast kräusen. It dropped out of suspension faster than DCY03.

DCY03: The “tan” isolate. This is significantly less funky, and a lot fruitier. It doesn’t produce much foaming during fermentation, and in fact may appear to be doing nothing at all. If shaken, though, there will be a lot of CO2 released. It did not produce a pellicle until several months after pitching and left the wort cloudy for about a month.


I hope to have more details soon, both from my beers but also from those of a few brewers I sent samples to. This was also the first time I’ve distributed anything so I kept it limited to four people. However, in the future I will release larger amounts to whomever wants it for a small fee to cover shipping and materials. Future announcements will be made on the blog and on Twitter.

PSA: Monitor your Yeast Bank

Much to my horror, recent attempts to revive four(!) different banked yeasts from my library have been unsuccessful. I fortunately have “backups” of the ones I consider most important, the “DCY” yeasts (which reminds me that I need to post an update about the DCambic yeasts, which are still fermenting one half-gallon each) and was able to work from that. But what I had considered my primary source of the two DCambic strains appears to be totally dead after just a few months.

I made a switch about six months ago to banking in isotonic NaCl solution after reading Samuel’s post on the matter. My previous attempts at reviving strains have succeeded, even with vials that were a few months older at the time. So, I’m suspecting operator error at this point and am very unwilling to cast this as a fault of the method. After all, Samuel used it for a while, and this is not a novel method. Possible problems include:

  • Improper solution: I used sea salt, because I couldn’t find a pure salt without additives. A lot contain minute amounts of dextrose, and I was trying to limit any sort of metabolism. Due to unknown levels of other salts in the mix it could require a different amount to be isotonic. I could also just have measured it wrong.
  • Storage temperatures at the back of the fridge were too low. Perhaps they dipped into freezing at times, and the small volume cooled off quickly.
  • I just moved a few weeks ago and somehow they got damaged in transit. Not likely, I don’t think, since I just moved within the same building and they were back in the fridge in just minutes.
  • Or, perhaps, I simply killed the sample while transferring into the salt solution. I had already revised my transfer method to avoid shocking yeast with a hot loop but it’s still a possibility.

The other two attempts were from slants. I have not had any issues in the past even with old slants, but these two for whatever reason opened slightly in storage and the tape around the cap wasn’t enough to hold the air in. They dried out as a result. I have successfully recovered yeast from such a slant before, but I am not so lucky this time–not yet, at least. As a last-ditch effort I will add sterile wort directly to the slant vial instead of picking off with a loop.

In either case, though, this highlights the need for monitoring things and perhaps something as simple as periodically tightening the caps on my vials. I may switch back to slants, at least temporarily, while I figure out what could be going wrong with the salt solution method.

Update: I was able to wake the two slanted strains up from their dead, dried, reddish state by adding sterile wort directly to the slant and shaking vigorously throughout the evening. In addition to having dried off, they were also slanted on media I made very early on that wasn’t ideal. I am surprised, frankly, that I was able to do this.

Update #2: Correcting update #1, it looks like only one of the slants actually yielded live yeast. As for the two DCambic yeasts from the NaCl storage medium, colonies finally are visible a week after plating. This is not unusual for Brettanomyces left at 20-22C/68-72F so this may have been much ado about nothing.

Sourdough Plating #3 and DCambic Update

This is the third and final attempt at plating out the whole sourdough starter. I did a few things differently:

  • I fed the three sourdough starters and let them ferment for 12 hours before starting. Last time two of the starters were right out of the fridge.
  • I put even less of each culture in the sterile water to dilute it
  • I used bromocresol green + potato dextrose plates, plus one potato lactose plate

The rye starter’s plate grew mold in a few days. It totally took over before I had a chance to see anything else growing. The other two starter’s plates plus the lactose plate (on which I plated the wheat starter) were totally fine. I let them grow out for 16 days.

Clockwise from top left: brown rice flour starter, wheat flour starter, wheat starter on potato lactose plate

The colonies look translucent and shiny from certain lighting angles, but they are off-white, smooth and round and actually matte. I did not see any flat colonies like last time. The weirdest surprise is the almost identical growth on the lactose plate. I am going to assume that this is due to whatever sugar was in the potato and not the lactose. My goal in doing so was to see if I could get the bacteria to grow on there and was not expecting the yeast to grow at all. You can see very tiny colonies on each of the plates between the large yeast colonies and smaller yeast colonies. They are transparent and shiny. I believe these are the lactic bacteria.

I think I may not have used enough bromocresol green, or the plates were old enough that it had already started to fade. I was expecting green colonies here, too, but they were white. On my first (recent) attempt I did see green colonies but the plates were also darker back then. The colonies looked quite similar between the two starters. This is a change from the first time I did this (earlier in the year) but could be because I am using a different medium.

I picked out a yeast colony from each and plated it on YEPM medium (recipe from jaapie’s site; the ingredients are not cheap or easy to find for hobbyists like me). In the past, the yeast I isolated from the wheat starter didn’t appear to ferment maltose whereas the rice starter’s did. We’ll see if these grow on it (from what I can tell, yeasts generally but not always can grow on something they can ferment and vice-versa).

On another note, I finally resumed the DCambic project from April. I had isolated a yeast from it back then but it appeared to not do much besides make lots of strong fruit esters. Notably, it didn’t seem to produce any alcohol. This time I instead went for growing it in liquid first instead of plating directly from the sample. I now seem to have two different types of yeast in liquid, judging from the two colors of sediment I see. A darker sediment formed first. I then started shaking to aerate the culture on a regular basis for a few days and then added this to more sterile wort, which I similarly shake on a regular basis. I am now seeing a lighter color sediment forming as well. I also see lots of evidence of fermentation, though I did not see any krausen at any point. I shook up the culture yesterday and streaked a sample on a PDA plate. I’m hopeful that this time I can get the yeast that produces the distinct Bretty aroma I can smell from the culture.

Sourdough Plating Results

These are the results after 5 days of growth.

Wheat flour starter (Carl’s Friends) on top. Bottom left is the brown rice flour starter (homegrown) and the right is the immature rye flour starter.

I am a bit surprised for a few reasons. First, from looking at them I was convinced that there was no yeast growth. At first I had all three of them in a big plastic bag, and the smell was not at all yeasty. It also looked like all of the colonies on the wheat and brown rice starters were very glossy like bacterial colonies. I had previously plated both of those on malt extract agar and the yeast colonies looked very different from each other, but definitely looked like yeast. I saw nothing of the sort here. The rye starter definitely had nothing that looked like yeast, but it is still young.

The first surprise came when I opened the tops to photograph them. I first took some of the colonies from the wheat starter plate and put them in a bit of light wort. But when I took a whiff both the wheat starter plate and rice starter plate smelled very yeasty. The smell from the bag I mentioned before was from the rye starter plate. It smelled not wholly unpleasant. I could pick out acetic acid but there were other things underneath. The colonies were tiny and slightly yellow. I’m not sure if they were heterofermentative LAB that produce acetic acid or some sort of enteric bacteria.

In the end I tossed out all the plates. But now looking at the picture above I can see that the colonies aren’t equally glossy. On malt extract agar (a long time ago) the wheat starter’s yeast were flat, dull and bigger than brewer’s yeast. That kind of matches some of the colonies I see in the picture. The rice starter’s yeast looked and behaved a lot like brewer’s yeast (I planned on trying it for brewing but never got around to it. I just made some small liquid cultures.) I don’t see anything of the sort here.

Later today I’m going for a third try on the bromocresol green and dextrose plates. Hopefully I can get some distinct yeast growth since LAB aren’t supposed to grow on it.

Sourdough Project False Starts

The sourdough stuff I promised in the last post hasn’t gone too well so far. I made the plates I mentioned but doing things with them has not gone well.

The idea is to do a few things:

  • Compare the organisms in three sourdough starters
  • Evaluate the properties of the yeasts I can isolate
  • Try to learn a thing or two about lactic bacteria
  • Try to cultivate a new starter (the third from above) from scratch, which I hadn’t done in a while

So the fourth bullet is sort of working, but not without some problems. I made an effort to use sanitized glasses and silverware to handle the starter so that I could be reasonably sure the organisms were coming from the flour and not from my environment. It took a long time but I can’t say whether that is due to that or some other reason. After a lot of strange stages (producing what I think is DMS, since it smelled a lot like cooked corn) it looks like it is becoming a viable starter. However, I messed up one day and stuck an unsanitized fork in there. It was just starting to turn around, but it sped up its turnaround after that. Having recently worked with other starters it could be I “contaminated” it with known-good organisms.

(As a side note, I always wonder if the number of environmental bugs is sufficient to overcome whatever is in the flour, but I suspect that is highly complicated and not subject to a simple yes/no assessment. I wish I had access to irradiated flour so I could put it to some sort of test.)

The first three require me to plate the starters. My first effort at this was destined to fail: I tried to streak after dipping a needle in the starters. I got some bugs from the early 2nd day stage of my new starter, some bacteria I can’t identify (yellow colonies on PDA). The other plates grew only a handful of yeast colonies but the plates grew tons of mold before I could do anything else with them. In any case, the bits of liquid dough along the streaks made it quite difficult to tell if things were growing at first.

I repeated the effort this unplanned long weekend, since I was home from the storm. I went with spreading a sample of sterile water that had a bit of starter dipped in it. After 24 hours there is no growth at 75F. We’ll see what happens in the next couple of days.