Decoding some White Labs strains from the Gallone et al. 2016 paper

NOTE: This post has been updated.

The paper on brewing yeast domestication authored by the Verstrepen lab and White Labs last year is a fantastic piece of work (i.e. Domestication and Divergence of Saccharomyces cerevisiae Beer Yeasts, Gallone et al. 2016, Cell). I’m sure most of you have seen it by now, so I won’t be going into any details about it. While the article and the work that has been done is amazing, there is one big negative about it: the strains used in the study have been coded, and there is no way of knowing what strain in the paper corresponds to what strain from White Labs or the Verstrepen lab. This means that the readers can’t really benefit from the huge amount of phenotypic data that the study generated. Wouldn’t be nice, for example, to know which White Labs strains are POF+, which ones can’t use maltotriose, which ones produce high concentrations of isoamyl acetate, or which ones sporulate easily?

There is one way of finding out what the coded strains are though, but unfortunately it isn’t completely straight-forward and at the moment it can only be done with a handful of strains. The authors have released the genome assemblies for each of the strains used in the study (available here: https://www.ncbi.nlm.nih.gov/bioproject/323691). The genomes of some White Labs strains have been sequenced in other studies, and performing a search for ‘WLP*’ in the NCBI database yields some hits (https://www.ncbi.nlm.nih.gov/biosample/?term=WLP*). Most of these Illumina reads are from a recent paper focusing on wine yeasts by Borneman et al. 2016. Luckily the strains weren’t coded in this study, and we know what White Labs strains these sequence reads are derived from.

What I then did, was download all of the raw sequence reads for any White Labs strains I could find. I then aligned them to a brewing yeast reference genome (VTT-A81062). After alignment, I looked for SNPs with FreeBayes, and used the resulting VCF files to create consensus sequences for each White Labs strain using BCFtools. I then used ParSNP to perform core genome alignment on these consensus sequences and all 157 of the assemblies from the Gallone et al. paper that I had obtained previously. ParSNP outputs a core genome phylogeny (generated with FastTree2) and a SNP matrix. I additionally produced a maximum likelihood phylogenetic tree using the SNP matrix and ExaML. To identify the White Labs strains from the coded strains in the Gallone et al. paper, I then looked for the closest hits in the phylogenetic tree (this was really obvious for most strains). In cases when there wasn’t one obvious hit, I also looked at Supplementary Table S1 in Gallone et al. 2016 for the reported origin and source of each strain.

Using this approach, I was able to identify eight White Labs strains from the set of coded strains. I know it is not much, but at least it is a good start, and it already gives some interesting and valuable information. First of all here are the results:

WLP_strainsThe question mark (?) after the ‘Code in paper’ means that there were two close hits, and I chose the match based on the reported origin and source as well. First of all, my suspicions regarding WLP099 were true (see this old post), with it being Beer033, a yeast grouping in the wine clade and unable to use maltotriose. Interestingly, WLP570 (Beer085) also seems to be unable to use maltotriose. This should be the ‘Duvel’ strain, and based on this information, it should only be able to produce bone-dry beers in worts supplemented with sugars (as Duvel is, if I’ve understood correctly). Another interesting observation we can make is that WLP705 appears to be Sake002. This is the only out of the seven Sake strains that wasn’t grouped in the Asian clade, but rather in the wine clade. Is this really a Sake yeast, or is it a mislabeled wine yeast? Anyways, there is a lot of interesting data to be extracted from these strains alone (esters, ethanol tolerance, etc.)!

While writing this blog post, I also noticed that the authors have released the Illumina reads to all the strains as well (in June 2017 apparently). This should allow me to confirm these results by looking at things such as chromosome copy numbers (through coverage) and SNP heterozygosity (the assemblies are haploid sequences). Unfortunately, I probably won’t have time to do that anytime soon though. But hopefully I can have a look at those later during the winter.

I will also be keeping an eye on the NCBI databases in case more White Labs strains are sequenced in the future!

75 thoughts on “Decoding some White Labs strains from the Gallone et al. 2016 paper

  1. suregork Post author

    Thanks, I wasn’t aware of this! Unfortunately it wasn’t that much help (nothing that contradicts my results here though). Annoyingly they’ve left out some of the interesting(/controversial) strains, such as WLP099 being a wine yeast. I’d also be interested in knowing the background of WLP644.

  2. Weehee

    I wonder if wlp570 is the primary strain for duvel or just a bottling strain as many Belgian breweries use wine yeast in their bottles.

    I have one bottle of duvel cellared and wlp570 in my fridge so I guess I could compare…

    I’ve been meaning to write about this but it’s kind of interesting that wlp028 is supposedly the McEwan’s strain – in which duvel originally isolated their house strain and one is pof+ and the other isn’t. I wonder if McEwan’s changed their yeast since it was banked by whitelabs/brewtek. Speaking of which there are many accounts of wlp028 making peaty smell (usually at higher gravities iirc) – perhaps this taste people are getting is not from the yeast but mishandled hops, for example I have a bag of EKG that must have been mishandled because it stinks of peat bog.

  3. suregork Post author

    Could very well be. Also, the reported origins of the White Labs strains are, as far as I’m aware, mostly speculation from the users and not anything that has been officially confirmed from White Labs.

  4. qq

    For those like me that didn’t read down far enough they’re hidden away on p18 and p22 :

    Beer Group 1:
    WLP001, WLP002, WLP004, WLP005, WLP007, WLP008, WLP011, WLP023, WLP025, WLP028, WLP029, WLP041, WLP090, WLP300, WLP320, WLP380, WLP400, WLP410, WLP510, WLP530, WLP540, WLP550 (core)
    WLP003, WLP006, WLP017, WLP022, WLP030, WLP039, WLP072 (vault)

    Beer Group 2:
    WLP545, WLP565, WLP566, WLP570 (core)
    WLP026, WLP037, WLP038, WLP585 (vault)

    Wine:
    WLP078 Neutral Grain Yeast

    Mixed:
    WLP036 Düsseldorf Alt Ale Yeast, WLP045 Scotch Whisky Yeast, WLP050 Tennessee Whiskey Yeast

    You can use that list with a bit of psychology to fill in some more gaps. For instance in the above table you have Beer057 is WLP 023 and Beer060. From the White Labs catalogue you can see that WLP025 and WLP026 were also used – wouldn’t you bet that they were Beer058 and Beer059? Sure enough Beer058 is a Beer1 yeast just like WLP025 (and on the Belgiumy end of the US/UK group, not surprising for a yeast from Southwold which faces Belgium over the North Sea), whereas Beer059 is in the middle of the bigger subgroup of Beer2, and WLP026 is allegedly Marstons which would fit with Manchester & Yorkshire Square in a northern English grouping in Beer2. Interestingly this sequence data has inspired Wiper & True to make a saison with Yorkshire Square.

    It also fits the run of 59-63 where the groupings (which are distinctive for their weirdness round there) match the numerical order of WLP026-38, as long as you leave out WLP030 Thames (on the WL list) and WLP029 Kolsch & WLP031 Klassic (“Young’s”) which aren’t on the list.

    However this would imply WLP002 is 45 rather than 50 – is there a chance that could be correct? It would then look rather neat the way that WLP001-8 match the groupings of 44-51, US/UK/BE-DE/UK/UK/UK/UK/US.

    There’s uncertainty between 52-56 as there’s six possible WLP yeasts for 5 code numbers, but only WLP011 and 017 are listed. This could be that because they already had one of WLP009-022 in the project under another name, or simply it was DOA or the sequence screwed up, there’s a few gaps later where they could have put it back in. Maybe people more familiar with the detailed history might be able to draw some conclusions based on close study of the dendrogram? But for instance it’s reassuring that 44 and 67 are adjacent, when I have them as Chico and San Diego.

    The other remaining uncertainty is in the Group 1 Belgians – 72-78 and 87 matching 8xWLP’s. Again it looks like something died or had bad sequence, so got added on as 87. That means we can’t be sure which of the eight yeast WLP300-540 are which code number (but we can guess 1 either way) – looking at the dendrogram I’m fairly happy it happened among the wheat beer yeast rather than the abbey yeasts.

  5. qq

    So here is my PROVISIONAL alignment of the codes with WLP yeast – this is just guesswork but could make confirmation easier through eg some basic PCR or something.

    44 1US WLP001 – California Ale
    45 1UK WLP002 – English Ale
    46 1BE WLP003 – German Ale II
    47 1UK WLP004 – Irish Ale
    48 1UK WLP005 – English Ale
    49 1UK WLP006 – Bedford British Ale
    50 1UK WLP007 – Dry English Ale
    51 1US WLP008 – East Coast Ale
    52 1US xWLP009 – Australian Ale
    xWLP011 – European Ale
    53 1UK xWLP013 – London Ale
    54 1UK xWLP017 – Whitbread II Ale
    55 1UK xWLP019 – California IV Ale
    56 1UK xWLP022 – Essex Ale
    57 1UK WLP023 – Burton Ale
    58 1US WLP025 – Southwold
    59 2UK Wlp026 – Premium Bitter Ale
    60 1UK WLP028 – Edinburgh Ale
    61 MIX WLP036 – Dusseldorf Alt Ale
    62 2UK WLP037 – Yorkshire Square Ale
    63 2UK WLP038 – Manchester
    64 1UK WLP039 – East Midlands Ale
    65 1US WLP041 – Pacific Ale
    66 1UK WLP072 – French Ale
    67 1US WLP090 -San Diego Super
    68 1US
    69 1UK ??this block & 79/82 has WLP030 Thames + one from WLP009-022??
    70 1UK
    71 1US
    xWLP300 – Hefeweizen Ale
    72 1BE xWLP320 – American Hefe Ale
    73 1BE xWLP380 – Hefeweizen Ale IV
    74 1BE xWLP400 – Belgian Wit Ale
    75 1BE xWLP410 – Belgian Wit II
    76 1BE WLP510 – Bastogne Belgian Ale
    77 1BE WLP530 – Abbey Ale
    78 1BE WLP540s – Abbey IV
    79 1UK
    80 2UK WLP545 – Belgian Strong Ale
    81 1BE WLP550 – Belgian Ale
    82 1US
    83 2UK WLP565 – Saison Ale
    84 2UK WLP566 – Belgian Saison II
    85 2BE WLP570 – Belgian Golden Ale
    86 2UK WLP585s – Belgian Saison III
    87 1BE ??The missing one of WLP300-410??
    88 WIN – any chance this could be WLP099??

    EDIT: Table removed as requested (updated table below)

  6. qq

    [please delete that previous table]
    I’ve now found Table S1 and used its “Origin” column to refine things a bit further. Turns out that WLP019 California IV is the one that is out of sequence, but it could be 65, 68,69 which are all listed as originating in California (the first two are Chico relatives although 69 is a UK Group1, related to Whitbread II. Is that WLP051 California V, aka BRY-97 Ballantine/Anchor???).

    I didn’t quite get the alignment right in the high 60’s/early 70’s before, but it’s easy with origins included. The only English beer for WLP030 Thames is 67, and I guess WLP029 Kolsch is 87. There’s not a second English beer slot free for WLP033 Klassic, on the assumption that it’s Young’s unless we extend the WL group out to the English beer at 90 or beyond.

    The Belgians sort themselves out once you extend them out to the UK Group 1 Trappist at 79 (Rochefort?), so this is what the list looks like now :

    BE044 1US WLP001 – California Ale
    BE045 1UK WLP002 – English Ale
    BE046 1BE WLP003 – German Ale II
    BE047 1UK WLP004 – Irish Ale
    BE048 1UK WLP005 – English Ale
    BE049 1UK WLP006 – Bedford British Ale
    BE050 1UK WLP007 – Dry English Ale
    BE051 1US WLP008 – East Coast Ale
    BE052 1US WLP009 – Australian Ale
    BE053 1UK WLP011 – European Ale
    BE054 1UK WLP013 – London Ale
    BE055 1UK WLP017 – Whitbread II Ale
    BE056 1UK WLP022 – Essex Ale
    BE057 1UK WLP023 – Burton Ale
    BE058 1US WLP025 – Southwold
    BE059 2UK Wlp026 – Premium Bitter Ale
    BE060 1UK WLP028 – Edinburgh Ale
    BE061 MIX WLP036 – Dusseldorf Alt Ale
    BE062 2UK WLP037 – Yorkshire Square Ale
    BE063 2UK WLP038 – Manchester
    BE064 1UK WLP039 – East Midlands Ale
    BE065 1US xWLP019 – California IV Ale (or 68/69??)
    BE066 1UK WLP041 – Pacific Ale
    BE067 1US “English beer” – ???WLP030 Thames???
    BE068 1US “California pale ale”
    BE069 1UK “California beer”
    BE070 1UK WLP072 – French Ale
    BE071 1US WLP090 -San Diego Super
    BE072 1BE WLP300 – Hefeweizen Ale
    BE073 1BE WLP320 – American Hefe Ale
    BE074 1BE WLP380 – Hefeweizen Ale IV
    BE075 1BE WLP400 – Belgian Wit Ale
    BE076 1BE WLP410 – Belgian Wit II
    BE077 1BE WLP510 – Bastogne Belgian Ale
    BE078 1BE WLP530 – Abbey Ale
    BE079 1UK WLP540 – Abbey IV
    BE080 2UK WLP545 – Belgian Strong Ale
    BE081 1BE WLP550 – Belgian Ale
    BE082 1US “Belgian beer”
    BE083 2UK WLP565 – Saison Ale
    BE084 2UK WLP566 – Belgian Saison II
    BE085 2BE WLP570 – Belgian Golden Ale
    BE086 2UK WLP585s – Belgian Saison III
    BE087 1BE “German lager” (but ale yeast?) – WLP029 Kolsch??
    BE088 WIN “German beer” (but wine yeast)

    I see 88 has its origins in German beer – but that might not necessarily exclude WLP099. Thomas Hardy’s was brewed with the normal Eldridge Pope yeast AIUI – but for the 6-month secondary fermentation in barrels they used a yeast they had got from Germany to make their (notoriously bad) Faust lager. So if White Labs or their forerunners had taken yeast from a bottle of TH, then they probably recovered the Faust yeast to turn into WLP099….

  7. qq

    Looking further at Table S1, the only Spirits yeast from Tennessee is SP009 so I guess that must be WLP050, which suggests that of the two Scottish spirits yeasts, SP008 is probably WLP045 and since little seems to be known about WLP078 Neutral Grain Yeast I suspect it must be SP011, a wine group yeast of unknown origin.

  8. qq

    Looking further at Table S1, the only Spirits yeast from Tennessee is SP009 so I guess that must be WLP050, which suggests that of the two Scottish spirits yeasts, SP008 is probably WLP045 and since little seems to be known about WLP078 Neutral Grain Yeast I suspect it must be SP011, a wine group yeast of unknown origin. So is SP010 a White Labs strain between WLP050 and WLP078? No obvious candidates for an Unknown origin.

  9. suregork Post author

    Wow, thank you so much for the comments and the great detective work!

    WLP002 could definitely be Beer045 instead of Beer050 (which in turn would make sense to be the closely related WLP007). This was one of the strains where the result was a bit uncertain.

    I’ve since posting this managed to identify two more strains, Beer072 as WLP300 and Beer083 as WLP565, and these fit in with your guesses.

    Regarding WLP099, I’m quite certain that it is Beer033. I’m not sure why it would be #033. Beer088 is also a close relative, but the problem is that this strain uses maltotriose, which WLP099 seems unable to. This suggests it can’t be Beer088. The origin of Beer033 also supports it being WLP099. But I’ll look into this in more detail!

    Again, very nice work! Would you mind if I repost your guessed strain table?

  10. qq

    That’s nice that my guesses on 72 and 83 match your answers doing it the hard way!!! 🙂 It has to be emphasised that this is all tentative at this stage, the only way to really confirm it would be to design PCRs that could distinguish them in the lab. WLP019/029/030 are the itches that really need scratching, they don’t fit in the main sequence and so I’ve just put them in the most appropriate slot in the WL block, but like 099 they may well be somewhere else.

    I presume that 099 is at Beer33 because they already had it in the lab, and the same could be true of WLP019/029/030 but there aren’t too many slots for Group 1’s of British or US origin, there’s lots of German lagers though (are they all cerevisae kolsch or is there pastorianus in there too??). It might be worth having a look through past papers from the lab to see if they have any obvious “favourites” that they’ve used in the past – how many Ukrainian and Bulgarian beer yeasts are they likely to have???

    I see eg 35-37, a Trappist beer from Belgium whose yeast has come from the UK, and I immediately think of Orval – interestingly they are not closely linked to 77, which I believe to be WLP510 which is traditionally linked to Orval. They are close to 79 =”WLP540″ linked to Rochefort, which I wasn’t aware had a British connection.

    However 31 is a Belgian Trappist that is close to 78, which I think is WLP530, linked to Westmalle.

    One mystery is that there’s no obvious WLP095, and one would expect them to do something with a member of the Conan family, outside the WL block there’s only Beer007 and 102 given as non-specific US beers, and they’re both US Group 1’s, when you’d expect Conan to be more obviously British. Maybe it’s listed as a British origin, it seems quite Yorkshire from what I can tell. I’ve seen a basic interdelta PCR that looked quite similar to Wyeast1968 (ie “WLP002/Fuller’s”), but you can’t rely on that. It might be Beer099 though.

    Yep, feel free to do what you want with the table – hopefully it will encourage people to play more with it. Could you delete the first version from Monday though, to avoid confusion?

  11. qq

    I’d vaguely heard about Chico/1056/WLP001 coming out of Ballantines, but I’m not sure I’d heard the full story. Interesting that it’s diploid (it’s GSY708 in Dunn & Sherlock 2008 doi/10.1101/gr.076075.108)

    However I’m a bit sceptical about BRY-97 being Conan. The history/geography might fit but I’ve never heard of BRY-97 fermenting anything like the profile of Conan. I know dry yeasts behave a bit differently, but this seems a bit of an ask.

    Also I’d invite you to look at this contribution to the Treehouse thread on HBT, where delta2/delta12 PCR of a Conan, WLP4000 Vermont Ale (#9) looks most like #5 1968 (WLP001/Fuller’s/”Beer044″) but does have some distinctive bands in common with #2 1272 (WLP051/BRY-97/”Beer068 or 69??”) and #3 1332(Hales/Gales)

    I must admit, I wish I had easy access to a PCR machine again! There’s all sorts of questions I want to ask, like what’s the composition of Nottingham dry yeast? It’s known to be a blend, and supposedly it is 70% “lager yeast”. I’d always assumed that was probably Lallemand’s Diamond Lager strain, which is from Germany and so presumably a Frohberg type lager yeast, which should be OK at ale temps. But conceivably that component could be BRY-96, an ale yeast that Ballantine’s used for lager. And then who knows, maybe one of the other components could be BRY-97?? It’d be interested to isolate some colonies from Nottingham and have a play – are there Frohberg colonies in there, and how do the ale ones compare with WLP039/”Beer064″ which supposedly shares its origin with dry Notty? It’s interesting that Lallemand have had a majority stake in Seibel since the early 2000’s.

  12. qq

    Don’t forget that multistrains are the norm in European ale breweries, so you can’t say there is just one “strain”. And that’s before you consider bottling yeasts. So even if the supposed history is correct, McEwan’s could have been brewed with a mix of WLP028 _and_ WLP570 (and others).

    “Peaty” notes sounds like phenolics to me…

    McEwan’s have had a tough time of it -the Export was one of the biggest beers in Britain in the 70’s, but suffered in the face of the popularity of lager, Scottish & Newcastle closed the brewery in 2005 and brewed it at their other breweries, then they were bought by Heineken in 2008, McEwan’s was sold to Charles Wells in 2011 and Marston’s bought Wells’ brewing interests earlier this year.

  13. Weehee

    Edsbeerblog streaked nottingham and corroborates what you say although he never tested them. The other option for conan is ringwood I heard, maybe one of the yeasts from the multiculture, too many posts to try to source that quote though so take it with a pinch of salt.

    https://www.homebrewtalk.com/forum/threads/achieving-a-silky-pillowy-creamy-mouthfeel-a-la-hill-farmstead.517721/page-17#post-7839270

    It’d be funny if conan was actually a scottish ale strain, there’s one old one which has a similar description (apricots). Although Noonan published his book in 93 not in the 80s.

    Yeah I sent a bunch of yeasts to that guy, I only read up until post 80ish. Did they actually prove it was blended yeast or is there any way of determining cross contamination?

  14. Weehee

    Very true! I wonder if they are related at all. There might have been a yeast change over time though. Yeah the old mcewans brewery is now flats and the storehouse is a school.

    It’s curious that higher gravities and temperatures seem to flush out phenols in UK strains while in wheat beer strains the opposite is true – iirc wheat beer strains have the same amount of phenols at any temperature but the ester/higher alcohol content alters perception of phenols – usually higher temperatures = more fruity notes. Even wlp099 seems to have that in the description: “As the gravity increases, some phenolic character is evident, followed by the winey-ness of beers over 16% ABV.”

  15. qq

    Edd was my source, the fact that the “lager” component of Nottingham grew at 37C for him suggests at the very least it’s quite a warm-adapted one, and certainly raises the question whether it’s not in fact something like Chico. It’s intriguing that one of your links suggests that Notty was a “grandparent” of Chico. I like it when brewing folklore and molecular evidence come together – qv the interdelta PCR that shows Windsor and Danstar ESB looking near-identical, and the comment in your HBA link that ESB is “sort of a spin off of Windsor”. Wouldn’t it be cute if Windsor was the ale component of Nottingham? 🙂

    Also interesting that Ed reports that “the Surebrew guy” found 5 strains in US-05. If Chico is diploid then it will be less stable, which also makes me wonder if Conan is diploid.

    I’ve seen Bierhaus claim before that Conan is a)NCYC1188 and b) this means it’s the “other half” of the Ringwood strain, but I’ve not seen anything else to support it. Conan does feel “northern English” – perhaps the closest I’ve drunk to it is 1469/Tim Taylor (but it’s definitely different). Ringwood’s yeast supposedly came from Hull originally, but I guess I’m just biased against the brewery as much as the yeast….

    The origins of the Ballantine yeast are intriguing – it’s known they had a big shakeup in the 1960s, so it could be as recent as that. Although MacKechnie was Scottish, he worked for Watney in London before he went to Canada (and everyone seems to think the Ballantine “lager” was inspired by Canadian beers) and Ballantine recruited him from London in 1933. London in the 1930s makes me think of the Mann yeast, which was widely traded at that time and had a good reputation. But Ballantine’s also had a German brewer working on the lager side before MacKechnie (and BRY-96/Chico/GSY708 is close to the ale parent of Saaz lagers, so maybe it has a central European origin), and a “master brewer from Burton” helping them with their Burton ale (and maybe IPA????) so it wouldn’t surprise me if BRY-97 had come from Burton originally.

  16. qq

    I’m not an expert on this kind of stuff but just trying to work it out from first principles, what if in POF+ strains there’s plenty of enzyme and substrate is limiting, which means it won’t be terribly temperature dependent. Imagine in POF- strains the main pathway is mutated but there is some residual phenolic conversion from secondary pathways, but enzyme is limiting. In which case you’d expect more enzyme activity at higher temperatures, and that might make the difference between undetectable phenolics at 16C and a faint whiff of them at 20C or whatever. Higher gravity means they’re being more stressed (they’re bathing in poison after all) so you could get all sorts of stress responses.

  17. notmyrealname

    First of all, I’d like to thank you for doing this excellent detective work. I’m sure there are valid intellectual property concerns why they just can’t come out and name the individual strains in the phylogeny, most likely the companies involved just don’t ever want their strains, named, catalogued and released to the public and prefer to keep that knowledge in house. So your investigative work here is much appreciated by us plebs who realise that we may never get a proper taxonomy of the strains sold to us. I would like to add my 2 cents to the discussion, if I may.

    Firstly, my empirical, therefore highly anecdotal, brewing evidence supports the conclusion that Rochefort is indeed an atypical yeast for brewing Belgian style beers. I’ve told those who’d listen that the Rochefort beer style, and the yeast, was of likely British origin for years now. Rochefort is made very much in the manner of English ales in the mid-late 1800s, brewed strong with loads of dark unrefined sugar and phenolic beer was rejected, so you needed a PAD1- yeast. Fruitiness yes, phenols a big no-no. In my several trips to the Lowlands over the years I’ve encountered brewers both big and small that told of similar tales. The small brewery DePrael (of Amsterdam) admitted once they used an english yeast to brew both their tripel and quadrupel. One thing these strains all have in common with typical English yeasts is suppressed (or absent) phenols and more abundant fusels and ethyl acetate, especially if you use sucrose. (see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4383656/pdf/yea0032-0281.pdf ). Rochefort’s (wlp540) metabolites are just so typically English and apparently n-propanol is a marker for domestication on that island.

    Secondly, I have serious reservations about the UCL article by Gallone (2016). The entire dataset is based on the supposition that the yeast strains involved in the study are pure species S. cerevisiae. The do not appear to have screened the 157 involved to ensure they were pure and I can tell from the inclusion of BE 31 and BE 77-78 (likely Chimay is at least one of these) and BE 30 (likely Westmalle or Orval) that have missed the fact that these 3 Trappist strains are proven to be hybrids of S.cerevisiae and S. kudriavzevii ( http://aem.asm.org/content/74/8/2314.full or https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3468397/ ) Both of those studies are from 10 years ago so the results are not new or earth shattering. My question to you guys and your more in-depth knowledge of genomics is this: Is the presence of complex interspecies-hybrids a problem for this study?

    I’ve read and re-read this article about a dozen times since it was released and even I contacted the authors about my concerns and posed them the exact same question, but have never heard back from them. I understand my background is in a completely different scientific field, but I have enough understanding to apply it to microbiology. I think the presence of a diploid set of S. cerevisiae chromosomes and a haploid set of S. kudriavzevii chromosomes in some of the strains present in this study warrants closer examination. Do you?

  18. Weehee

    Awesome stuff. This is going to end up going from a paper about the domestication of S cerevisiae to the story about how it spread! Definitely deserves a write up, maybe even a tree diagram of sorts with different possibilities.

  19. Weehee

    I just had a thought, if there are/was 5 strains in us-05 I hope it doesn’t compromise any of the studies which used it…

    I read on the HBT white labs lawsuit thread that there’s allegedly some london breweries with cross contamination of diastaticus from fermentis now, I hope have measures in place when making their new product…

    Also, apparently 50/50 Nottingham/Windsor is supposed to be very good for a multistrain beer.

  20. qq

    I’ve remembered now – the logic that says Conan is the other half of Ringwood’s production yeast goes :

    1) Conan is allegedly NCYC 1188 (evidence??)
    2) NCYC 1187 and 1188 are from the same brewery (quite possible but evidence??)
    3) NCYC 1187 is allegedly Wyeast 1187 (is there evidence of Wyeast using NCYC #’s in this way??)
    4) Wyeast refer to 1187 as “Ringwood Ale” – but what’s the provenance?

    So a whole lot of coulda-woulda-shoulda, which rests in particular on Wyeast translating NCYC numbers directly. I don’t see much reason for confidence that Conan came from Ringwood.

    Going back to the Californias. An obvious assumption is that the three “California beer” strains in the middle of the WL block – BE065/68/69 – correspond to WLP019 California IV Ale (“Moderately clean strain with a low ester profile and less sulfur production than WLP051”), WLP051 California V Ale (“fruiter than WLP001 and slightly more flocculent. Attenuation is lower…than WLP001”) and WLP076 Old Sonoma Ale (“a traditional British-style yeast”). That sounds like two US strains and a UK strain – and sure enough that’s what BE065/68/69 are. So tentatively, I propose :

    BE065 1US WLP019 California IV (source??? “Cleaner 051”)
    BE068 1US WLP051 California V (Anchor/BRY-97/Ballantine ale)
    BE069 1UK WLP076 Old Sonoma (New Albion)

    I don’t know why 65-69 are out of order. It’s interesting that most of them are from California, so I wonder if they are actually the version from “local” breweries that use WL for banking etc, so they have some other code number whilst being almost but not quite the same as the strains we know as WLP019/051/076.

    Intriguingly BE069 is closest to BE079, putatively WLP540 “Rochefort”. That would kinda fit what we know of the Heath Robinson approach Jack McAuliffe was forced to take when he create New Albion, generally regarded as the first “new” US “craft” brewery – he adapted dairy equipment, he had to take whatever hops the macro brewers wouldn’t buy and so on. In that context, you would expect his yeast to come from whatever bottle-fermented beer he could get his hands on, and Belgian beers would be have been one of the few at that time.

  21. suregork Post author

    Thanks for your comment and input! Your point about hybrids is a very good one. As you say, many of the trappist strains are known to be S. cerevisiae x S. kudriavzevii hybrids. Of these, one would certainly expect to see some of the strains among the studied strains (e.g. I’ve understood WLP530 is the Westmalle strain and WLP500 the Chimay strain). I’m not sure how and if they’ve taken into account that the strains are hybrids. I’m no expert in the area, but from what I’ve understood, if you perform de novo assembly of sequencing reads from a hybrid strain, you would expect to see an increased size in the genome assembly. However, all the assemblies included in the study are, according to Table S1, around 12Mb (i.e. the size of a haploid S. cerevisiae genome). My guess is that this either means that a) there are no hybrid strains included in the study, or b) they have only retained the S. cerevisiae sub-genome of these strains for the analysis (if this is the case, I’m not sure how they’ve dealt with chimeric chromosomes and why they haven’t mentioned in the methods). B) is something similar to what has been done with the lager yeast strains in this very similar study: http://www.sciencedirect.com/science/article/pii/S0960982216309848. Now that the raw sequencing reads have been made public, I could go and have a look at some of the suspected trappist strains, and see how well they map to a S. kudriavzevii reference.

  22. Weehee

    Wyeast are known to hat tip their numbers according to this:

    https://www.homebrewersassociation.org/forum/index.php?topic=23662.0

    Conan is also called ypb 1188.

    Ncyc has been known to deposit their strains in sequential format (for example the labatt strain): http://www.jimsbeerkit.co.uk/forum/viewtopic.php?f=12&t=70705&start=15

    Here’s a thread with some of the ncyc strains for brewing: http://www.jimsbeerkit.co.uk/forum/viewtopic.php?f=12&t=58625

    Interestingly 1187 was mentioned it might be a hybrid – the british ale yeast pdf posted below by notmyrealname indicates that they tested a british strain for this hybrid but determined it wasn’t one. So we might be able to get information about 1187 from that paper if they are the same yeast (ringwood).

  23. suregork Post author

    And just a quick follow up. The Borneman et al. 2016 study (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4825664/) mentioned in the blog post also included WLP500 (the Chimay strain). They show, as we already know from previous studies, that it is an interspecific hybrid between S. cerevisiae and S. kudriavzevii. Interestingly, according to the information in the White Labs catalogue, WLP500 is not among the 157 strains in the Gallone et al. paper. This would suggest that the authors have left out any interspecific hybrids, and all the strains are ‘pure’ S. cerevisiae (as suggested by the assembly sizes). Is WLP530 then not a S. cerevisiae x S. kudriavzevii hybrid, even though the Westmalle strain is supposed to be one (http://aem.asm.org/content/74/8/2314.full)?

  24. suregork Post author

    Another potential ID. The Borneman et al. paper (mentioned in the blog post above) included WLP800 as well. Interestingly, they reveal that it is actually a S. cerevisiae strain, not a S. pastorianus strain. I had initially not seen this, so didn’t include the sequencing reads from WLP800 in my analysis. I repeated the analysis with this strain, and the closest hit for WLP800 is Beer087. The reported origin is Germany, lager beer, so it does make sense. Interesting if WLP800 isn’t actually a lager strain.

  25. qq

    Great stuff @notmyrealname, it’s always nice when DNA says something a bit surprising (“BE079 is a Trappist with British origins”) that matches the story from anecdote (aka hard-won experience!).

    I wouldn’t sweat the kudriavzevii thing, it’s like students of human ethnicity worrying about gorillas – they’re working at a level where “foreign” is easily recognised. Mice and rats have about 85% DNA the same, I don’t know yeast genomes well but I’d guess kudriavzevii is about 95% the same as cerevisiae – but even if it was 99% the same, that would still mean that 1 in 100 DNA “letters” were different – in a genome of ~12,000,000. So you can resolve even quite small bits of kudriavzevii DNA – as Borneman et al 2016 did. Any fragments of foreign DNA will soon be obvious once you try to construct a family tree as they will segregate well away from the cerevisiae subgroups.

    In fact, Gallone et al flag up yeasts whose DNA is a mosaic of different cerevisiae subgroups, and discuss one in particular. The German hefeweizen yeast seem to be group 1 beer yeast (ie well-domesticated, clean, POF-) that have taken on ~10% of a wine group genome through hybridisation – including the bit with the “phenolics gene”.

    As far as the Gonzalez et al 2008 paper – don’t forget that just because eg MUCL 20488 was isolated from a Chimay beer, and WLP500 was _allegedly_ isolated from a Chimay beer, you don’t know that MUCL 20488 and WLP500 are the same yeast strain. Beer Decoded have found 29 different “strains” from over 12 species in Chimay Red Cap but only 11 “strains” in Chimay Tripel. So it looks like there’s a <10% chance that MUCL 20488 and WLP500 are the same thing.

  26. qq

    A classic example of the dog that didn’t bark – you’re quite right, you’d expect WLP800 (supposedly from Urquell) to have some eubayanus in it, it should look more like WLP862 Cry Havoc (interesting that there’s a bit of uvarum in the latter as well!).

    On the other hand, WLP800 does have a reputation for top-cropping and generally being a bit of a weird lager yeast. It’s also notable that Wyeast make a point of marketing 2001 as Pilsner Urquell H-Strain, as though they’ve got it direct from the brewery – I wonder if it’s the same or whether WLP800 is some kind of contaminant or cold-adapted bottling yeast? We know Urquell bought their yeast from Germany, maybe it was a kolsch yeast????

    There remains the problem of WLP029 Kolsch (“Fruh”). There’s a couple of German yeasts from “beer” up at BE088 (1UK) and BE089 (wine) but I’d guess it’s the German ale yeast at BE008 – presumably the lab already had it. I’d imagine that Leuven have close links with Weihenstephan and that yeast people on the Continent have a special fascination for kolsch, so you’d expect them to have it already. I’ve seen it suggested that WLP029 is Weihenstephan W-177 but don’t know how reliable that is, but you’d kinda expect Leuven to have access to W-177.

  27. Weehee

    WLP862’s source was apparently a budweiser keg.
    https://youtu.be/4NaNAfrgXJg?t=8m30s

    WLP800 could be from a different brewery or maybe even part of a mixed culture. Might be worth looking at what breweries open ferment pils if it has good top cropping characteristics.

    If you search in that jimsbeerkit ncyc listing there are quite a few cerevisiae 100% matches that come from lager breweries.

  28. qq

    It’s still a bit of a stretch though, isn’t it? It’s hardly proof that Wyeast 1187 = NCYC1187.

    As for Conan, I think you’re thinking of it’s “official” name, VPB-1188. Given that Greg Noonan opened the Vermont Pub & Brewery in November 1988, that seems one heck of a coincidence. I’d guess that the name merely reflects the place and time of the first batch made with what we now know as Conan. John Kimmich talks a bit about it here :
    https://www.youtube.com/watch?v=LdfySDN2mF0&t=38m04s
    He describes it as “the yeast that Greg brought back from England, years ago…that was Greg’s discovery, he was in England drinking beers way back when, he found that yeast, he started using it”. It’s possible that Noonan got it from the NCYC but it sounds more likely that he got it direct from a brewery or cask. The way the delta2/delta12 PCR of WLP4000 looks sooo similar to 1968, they’ve got to be first cousins at most.

  29. qq

    Good find on WLP862, I knew it came from Charlie but hadn’t heard the origin story.

    As for open-fermented pils – well there is one brewery that’s famous for it, namely Urquell :
    http://www.sciencemag.org/news/2017/07/microbe-new-science-found-self-fermented-beer

    but after the fall of Communism they invested a lot of money in conicals and I think the open fermenters just get used for a few specials now. But it’s plausible that the WL version was harvested from pre-conical Urquell.

  30. qq

    The history gets a bit complicated, but folk in Hull regard it as the capital of the East Riding of Yorkshire, so you don’t need to invoke Halifax in order to call the Ringwood yeast a Yorkshire yeast. But as Gallone demonstrates, you can get saison-like yeasts turning up in Yorkshire and British yeasts turning up in Trappist monasteries, so I wouldn”t put toooo much weight on the idea of “Yorkshire” yeasts per se.

    The sequence of “Ringwood” is irrelevant, if by “Ringwood” you mean Wyeast 1187 as it’s a different strain to Conan even if you believe the story that 1187 and Conan are part of the same multistrain.

    But I’ve not seen any gels from 1187 – it wasn’t used by isomerization over on HBT for instance. But Wyeast1187 is meant to be the same as WLP005, which I posit is Beer048 in this study, so someone could do a virtual PCR on the Beer048 genome…

  31. Weehee

    Nice, thanks.

    Yeah 1188 probably isn’t conan, here’s some interesting information on it though. This study was cited in the British ale yeast paper from below. It also has a few quotes about whitbred B.

    “Similar groups could be distinguished amongst the yeasts which form head. Thus cells of number of yeasts (group D) passed into yeast head early in the fermentation, leaving very little yeast in suspension. Examples are strains 1006, 1122 and 1188. This groups is equivalent to that in which the yeast sedimented very early (group A), but the consequences differ. Because yeast in head was not in contact with wort, fermentation ceased, in contrast to the case with sedimented yeasts. If these exceptionally strong head-forming yeasts were returned to suspension from time to time, fermentation continued.”

    Sounds like it could be a strain which needs lots of rousing. 1006 was in the White labs study later.

    http://onlinelibrary.wiley.com/doi/10.1002/j.2050-0416.1969.tb03233.x/full

    There’s more on whitbred b:

    “First, some yeasts (group A) sedimented abnormally early as result of the early development of flocculation. Examples are strains 1026 and 1234. Although the worts became clear in consequence, fermentation continued at normal rate as result of the activity of the sedimented yeast. It is noteworthy that with such yeast there is then substantial difference between the specific gravity of the fermenting liquid immediately above the yeast and that higher in the tube. This difference amounts to degrees of gravity at the most active stage of fermentation but eventually disappears.”

    I wonder if they have the other 4 yeasts in their bank. I’m guessing they do just not for commercial purpose – it looks like untappd shows they brewed 4 beers with 1006 1681 1187 1228 but not 1026.

    Wyeast have a cask beer yeast called 1026 but most people think 1098 is whitbred b and 1026 is from oranjebloom – I wonder if the source is Shepherd’s neame because they make that now? Then again it is called cask ale, apparently Marston’s is the largest producer of that and wlp026 is allegedly marstons, which also links to 1[026]. I really have to try using this Marston’s pedigree yeast. I’ve only seen one report from that yeastwhisperer guy using the ncyc oranjeboom isolated ale strain and they said it produced too much isoamyl acetate. Brewlabs has an alleged Marston’s strain which is POF+ and produces sulphur (wyeast 1026 is sulphur prone when stressed apparently) and a moderate ester producer.

  32. qq

    NCYC 1026 is definitely part of the Whitbread B family – and is generally linked to WLP007.

    The first time I saw that link between Wyeast 1026 and Oranjeboom, I assumed it must be a Shepherd Neame connection, as you say they have been contract-brewing Oranjeboom for years now. I guess it’s partly to do with the amount of contract brewing they do but Sheps are quite promiscuous with their yeast, they use a lot of different ones.

    The majority of traditional British beer is cask ale, so I wouldn’t use that to link 1026 to Marstons. But Burton beers are notorious for a whiff of sulphur – the high sulphate levels in the water obviously help but it wouldn’t surprise me if the yeast contribute to the “Burton snatch” as it’s known. BE059 is a Group 2 yeast (ie less domesticated and POF+) and I posit that it is WLP026 Premium Bitter, supposedly linked to Marstons so that fits.

  33. suregork Post author

    I just had a look at the sequencing data from some of the Trappist strains: Beer031, Beer077, Beer078 and Beer081. I aligned the sequencing reads to a concatenated reference genome consisting of all the Saccharomyces species, similarly as to what had been done in the Borneman paper, in order to identify whether they are hybrids. All four strains appear to be pure S. cerevisiae. This makes me believe that the authors of the Gallone paper have excluded any interspecies hybrids from the study. This would also explain why WLP500 seems to be lacking (according to the White Labs catalogue), as this has been shown to be a S. cerevisiae x S. kudriavzevii hybrid. If Beer078 really is WLP530, it is interesting to note that it appears to be a different Westmalle isolate than the one determined to be a hybrid in the studies above.

  34. qq

    Just poking around, there’s all sorts of cuteness comes out of this alignment. In the heart of the Group 1 UK’s is a small “European” subgroup, consisting of WLP004 “Guinness”, WLP011 European Ale (“Wissenschaftliche Station #338”) and WLP072 French Ale (BE047/53/70). That’s interesting as Monerawela et al 2015 found that the closest sequenced relative to the ale parent of Frohberg lager yeasts was a Guinness strain, (they call it NCYC1511, I think they mean NCYC727, Gilliland 1511). Maybe WLP011 is closer still?

    It would also be interesting to see what the closest relative of the ale parent of Saaz lager yeast is – how perfect would it be if it was WLP800/BE087?! The closest sequence Monerawela et al could find was Foster’s O – a cerevisiae from a brewery known for lager. Does anyone know what that’s related to?

    On closer inspection I get even more frustrated about BE067. Not only is it the only one out of BE044-87 that we don’t have a fairly solid guess for, it occupies a key position as the only yeast from England in the middle of the Group 1 US beer family and so a vital part of the question of where the clean US yeasts came from.

    The only strain we definitely don’t have a home for is WLP030 Thames Valley Ale, which was banked in 2001 and is described as “Very flocculant strain for all things English. Great for porters, stouts and ESBs. Lower ester production than most English strains but creates a bigger mouthfeel than most cleaner strains” – and relatively high attenuation at 72-78%. Maybe it was harvested from an ancient bottle of London porter, but the only current stout/porters I can think of from SE England in 2001 would have been Fuller’s and Mackeson. Mackeson could be a good fit – milk stout is a quirky style that people would want to preserve, it was massive in the mid-20th century and was even briefly brewed in the US. 2001 was a notable year for it as well – it was the year that Whitbread (including Mackeson) were bought by Interbrew (now AB Inbev). It’d be interesting if all the US strains originated in milk stout! Maybe WLP030 was rescued from an old bottle though.

    Another possibility might be WLP066 London Fog (??Mann’s??) which is a better fit numerically but is so new that it’s not even in the current catalogue so would probably have not had a number at the time of this study, maybe it snuck in at the end as BE090/99/100/101???

    One other thing I’ve noticed – if you look at the organism details for the genomes, all the beer ones at least have organism accession dates of 1950 (up to Beer043) or 1995 (Beer044 onwards). It’s like all their beer yeasts came from one of two sources – ???the Leuven bank and White Labs????

  35. suregork Post author

    I’ve been planning to check where lager yeasts fit on the tree for a while now, and your comment pushed me to start some analysis (I’ve done some related work with a ‘kveik’ strain that turned out to be a S. cerevisiae x S. uvarum hybrid). We’ve recently sequenced one of our model Frohberg lager yeast strains we work with at our lab (VTT-A83015), and I’ve now ‘extracted’ the S. cerevisiae portion of the sequence, and am repeating the analysis as described in the blog post. I’ll post the results tomorrow hopefully. It will be interesting to see if it fits in around WLP004.

    Yes, it would be nice to get a better idea of the identity of Beer067. I also wonder what the one Belgium-based strain Beer082 in the US clade is. Could it maybe be WLP515 Antwerp Ale, which is described as a very clean ale yeast? I don’t know what the connection to the USA would be though.

    Regarding the last paragraph in your comment: Going back to the ‘Methods’-section of the actual paper, you can see that strains Beer044-Beer102, Lab002, Spirits008-011, Wild001-007 and Wine019 have been sequenced at Illumina in San Diego. These are almost certainly White Labs strains, or samples provided by White Labs. I’m also guessing that one of the California strains in the end (BE090/99/100/101) is probably Wyeast’s 1056 that they’ve included in the study. This assumption is based on the quotes in this article: http://allaboutbeer.com/article/the-family-tree-of-yeast/
    This means that there might also be other non-White Labs yeasts in the end of the BeerXXX sample list.

  36. suregork Post author

    Here is the tree with the ‘A15’ Frohberg-type lager yeast (in red):

    tree

    As can be seen, the A15 strain breaks off from the ‘Beer 1’ clade at quite an early stage, and there are no close relatives among the studied strains. I guess this is quite expected, assuming the lager hybrid was formed several hundred years ago.

  37. qq

    Yep, I’d come to the conclusion that BE082 was probably WLP515 “de Koninck” – numerically it could conceivably have come out of WLP560, a limited-edition recreation of the Dupont blend, but I don’t think so. WLP515 is noted for its cleanness – and de Koninck are famous for serving the dregs in a shot glass, which you wouldn’t want to do with a traditional Belgian yeast!

    It’s interesting he says 10 of the sequences are from strains used for lager production, as only 9 are annotated as such in the list, I imagine one “German beer” may be a bock or something. BE094 is a Whitbread B derivative being used for lager in Germany, I’m more interested in the BE/DE Group 1 BE041/43/87 – the first in the kolsch group, the second closest to Achouffe, the third possibly Urquell as above. But that kind of thing is a good place to be finding an extra-clean strain, which could have made its way to the coast of Belgium. From there you’ve got all sorts of routes to the UK – the Hanseatic League and all the historic trading relationships, mass migration like the Huguenots, and more recently the huge logistics of the Napoleonic wars and World War I.

    If BE067 really is a Mackeson strain, then you have the intriguing prospect of three old breweries in port towns at the base of the US/UK clade – Mackeson in Kent who date back to 1669, de Koninck in Antwerp founded in 1833 and Adnams in Suffolk in 1872. I really want Mackeson to be part of it as the story just works – they were bought by Whitbread of London in 1929 (a company with sophisticated lab facilities and interest in yeast banking), and MacKechnie was recruited by Ballantine’s in 1933 after a spell of consultancy in London.

    Nice one with A15 – as you say, it’s just where you’d expect. I’m even more curious about the Saaz parent now!

    That article’s interesting, if I was doing a big sequencing project like this I would allow myself a few “play” ones. Thinking what I’d do – Conan would be definitely one of them. If we take it from isomerization’s PCR that it’s closely related to 1968/WLP002/BE045 then my guess would be WLP095 being BE090.

    Given the close relationship with the Yeast Bay, I bet there’s at least one of their strains in there – BE102 might be their Midwestern strain from Bell’s?

    I think BE098 must be Wyeast 1056 as the only one that explicitly mentions California – and it’s not great surprise that it’s not identical to WLP001, there’s been various experiments that have come to that conclusion. What other Wyeast strains would White Labs be interested in? Ones that WL don’t have an exact match for, particularly something like 1318 which is getting huge attention as the “new Conan” for New England IPAs. I’d fancy 1318 as one of BE099/100/101, which are close to Whitbread B, “Nottingham” and Ringwood respectively.

  38. Weehee

    It’s funny I was going to bid on a Mackeson’s bottle about a month back, looked from at least 1950…starting bid was too pricey imo.

    I was thinking you could see if T-58 has kudriavzveii in it then that’ll pretty much confirm it’s a trappist yeast.

  39. qq

    ” Beer044-Beer102, Lab002, Spirits008-011, Wild001-007 and Wine019 have been sequenced at Illumina in San Diego.”

    I think that means we can be fairly confident about the following :
    SP008 MIX White Labs WLP045 – Scotch Whisky
    SP009 MIX White Labs WLP050 – Tennessee Whiskey
    SP011 WIN White Labs WLP078 – Neutral Grain Yeast

    They’re all kinda weird ones, but they fit. Interesting how the two Scottish spirits yeasts are together in a weird hybrid group of their own – I can imagine that the whisky yeasts have been pretty much isolated from other parts of Europe for the last 200 years.

    That leaves SP010 which is a real freaky wild-looking thing, I guess it must be either WLP065 – American Whiskey or WLP070 – Bourbon and the other of those was already in the Leuven bank and was one of the early spirit sequences, I don’t know anything about spirit yeasts to say which is a mainstream one.

    Presumably European labs have generated most of the wine strains so White Labs didn’t have muchy to contribute there. So WI019 must be something they’ve collected themselves – and its position on the fringes of the Belgian Group 2’s suggests somewhere out of the way, outside Belgium. I wonder if it’s a component of WLP611 New Nordic Ale??? “Isolated from spontaneously fermented apples on a remote island off the coast of Denmark. This culture is a unique blend of three yeast strains, two belonging to Saccharomyces cerevisiae and one Torulaspora delbrueckii. Although originally thriving in the simple sugar fermentations such as wine and cider, this blend ferments maltose as well. This blend has a characteristic aroma profile that resemble classic styles such as Belgian saison or German hefeweizen, especially at higher temperatures.”

  40. notmyrealname

    The reason I brought this all up is that after several emails to the members of the group responsible for the paper, there was simply no response.

    This silence I took as an admission of guilt that they in no way made reference to how they arrived at the conclusion that the species in their study were pure S. cerevisiae. It’s either a glaring error in study design ( I don’t think it was), or a serious omission of the study parameters in the body of the text, i.e there should be some reference to whether they tested for pure species or how possible hybrids were dealt with. Given that the Saccharomyces species complex is so prone to hybridisation and even lateral gene transfer, I would have surmised that it was somewhat important to highlight the steps taken to ensure the members of such a broad study (157) were “pure”. I’ve read their methods section about 5 times now. (http://www.cell.com/action/showMethods?pii=S0092-8674%2816%2931071-6)

    It appears they simply lifted 157 strains from 2 banks and relied upon them for identification to which they said , “Here, these are S.cerevisiae”. Are they correctly identified as such by the banks? If so then I have no concern. But as many are discovering, a lot of strains are stored under incorrect taxonomy. I certainly wouldn’t trust white labs to have their commercial strains identified correctly, but I would have thought that UCL/KUL would have better control over their internal collection.

    All in all, I will remain suspicious of their entire dataset unless it can be explained in much better detail or it can be repeated by another group. I have serious issues with much of the research being pumped out on brewing microbiology currently. There are glaring omissions in logic and no repeatability studies are done to confirm or refute the findings of what are simply “one interpretation of data produced in our lab”. I’ve been in the science arena for near on 3 decades and I find the overall drop in standards these days alarming, particularly the “plot-holes” one can find in most papers released. I blame publish or perish for this race to the bottom we are seeing.

    As timely evidence of this I actually bring up one of the articles you guys posted here: https://f1000research.com/articles/6-1676

    “ However, we acknowledge that the results may contain false positives as no quality trimming of the sequencing reads was performed. “

    If I helped crowd-fund this I would ask for my money back. It’s lazy science to not eliminate sources of bias as easily identified as this. That entire studies usefulness is called into question. Was S. kudriavzevii really present as a whole species or simply as part of hybrid genomes? As far as I know S. kudriavzevii has never been isolated in the brewing environment, only from the wild.

    Anyway, on a much lighter note I am enjoying the conjecture of beer strains that you guys are discussing. As an avid home maltster/brewer who has been serial re-pitching for a long time, I’m fascinated with the obvious genetic drift one can observe after just 20-30 brews of a single commercial yeast. I’m also fascinated to the extent of obsession with producing beer with a yeast-driven profile. Much of the beer-flavoured hop water that’s getting pumped out by my local brewers is very disheartening to me. “Viva” esters, fusels and phenols I say.

  41. suregork Post author

    I agree with you completely regarding the lack of methodology used to ensure the strains were ‘pure’ S. cerevisiae. You can find a presentation on this study given by Kevin Verstrepen on YouTube, and in the slides he shows an extended tree with lager yeasts included. This makes me think they were aware of the hybrid problem, and that the 157 strains actually are S. cerevisiae (this is backed up by the genome assembly sizes). However, as you say, they should definitely have included methodology and data showing how they determined this. You might have seen this very similar study by an international group of yeast researchers that reached more or less the same results with a different set of yeast strains: http://www.cell.com/current-biology/fulltext/S0960-9822(16)30984-8?_returnURL=http%3A%2F%2Flinkinghub.elsevier.com%2Fretrieve%2Fpii%2FS0960982216309848%3Fshowall%3Dtrue
    They at least recognize the hybrid problem regarding lager yeasts, but I’m not sure if they’ve tested contributions from other Saccharomyces species.

    As you say, the F1000 article has some problems. I’m actually the 3rd reviewer that brought up these issues and I haven’t accepted the paper yet. At the moment the results are filled with false positives, because the authors haven’t removed poor quality sequencing data. I did my own analysis on several samples using the raw data available, and these results were much less dramatic: mostly just one or two species (S. cerevisiae and/or B. bruxellensis) in each beer.

    Thank you for your comment and contribution, I’m very happy to see so much response from other people!

  42. qq

    “It appears they simply lifted 157 strains from 2 banks and relied upon them for identification to which they said , “Here, these are S.cerevisiae”. Are they correctly identified as such by the banks?”

    It doesn’t matter what the banks say, it’s irrelevant. They’re taking 157 organisms – it could be rattlesnakes, it could be dolphins – and sequencing the entire genome of all of them. For these guys sequencing a genome is like brewing 5 gallons of beer, it’s completely routine. Once you’ve got the sequence, then you can compare them against the sequences from model strains – and it will pretty soon become apparent that the DNA hasn’t come from dolphins or rattlesnakes, but yeast. And as I said above, you have the resolution to determine that individual chromosomes have fragments of eg cerevisiae fused to kudriavzevii fused to more cerevisiae or whatever. As long as your species identification is right for your reference strains then you can be pretty confident that yes, they are all cerevisiae because that’s what the DNA is telling you, not what some yeast bank is telling you.

    They will have sequenced more than 157 strains, this is just a selection of the cerevisiae ones – as suregork says they’ve now expanded the dendrogram to include lager strains, see here : https://youtu.be/E6qBnBQuWF4?t=30m20s (you’ll enjoy the whole hour’s lecture, it’s impressive the breadth of what they’re trying to achieve – even if they are Inbev’s bitches…:-)) ) Saaz cluster on the right of the mixed clade, Frohberg like A15 are on the left of the Group 1 beer clade – not inconsistent with the possibility that they’re hybrids with an alt yeast and a kolsch yeast respectively.

    As for the Beer Decoded paper – ouch, I’m amazed they’re not cleaning up the data more. Although to be fair they have very different aims, they’re looking to identify a “zoo” of beasties from the “wild” when they know that some beers can contain many different species/genera/kingdoms, that kind of fishing expedition will always set its filters wide to start with and then draw them tight. But one would hope that there would have been some basic quality control before anything reached the public domain let alone a paper.

    Still, I’m not sure it’s particularly useful to compare something like that with something on the scale of what they’re doing in Leuven. In particular Beer Decoded are not really geared up to detect mosaics of different species AIUI, whereas such mosaics are really obvious when you’re dealing with entire genomes.

  43. qq

    Looking at the “wild” yeast, WL005-7 came from Denmark and were sequenced in San Diego, so they look better candidates for parts of WLP611 (and others from the same sampling trip?)

    So WI019 has to go back on the pile, I’ve no idea – it’s interesting that it’s somewhat related to a Bulgarian beer strain.

    There’s a great example of how extremely close genomes don’t mean close phenotypes at one end of Group 2, there’s three “Dupont” strains – BE034/83/84. Sadly it looks like the former had a duff run on the GCMS (it’s not the only one where they’ve left that kind of thing in the data rather than doing another run) but they’re all very different.

    Another intriguing one is the pair of highly-related Group 2’s BE039/40, one from a Czech lager brewery, one from a US lager brewery. Hmm, which US brewery has particularly close historical links to the Czechs? And how come they are using a POF- relative of BE011, a POF- abbey beer, and BE085, a POF+ strong ale from Belgium which may be “Duvel”?

  44. qq

    Having read a bit more, they seem to be relying heavily on calibrating their molecular clock on the colonisation of the US as “ground zero” for the US clade. Then they’re supporting it with the new lager sequences as they tie the lager hybridisation events to the introduction of the Reinheitsgebot in one small part of what is now Germany. However, if this data tells us anything is that you don’t need hybrid yeast to make lager, even in the modern world.

    I don’t know if there’s any devil in the details but I still think it’s possible that Ground Zero for the US clade is Prohibition or rather, the arrival of a European consultant at Ballantine’s. In this view, the likes of WLP009 Australian and BE066/WLP041 “Redhook” are more typical of the yeasts brought over by the colonists and were the “Neanderthals” that were largely swept away by the “H. sapiens” from “Antwerp” or wherever.

    On the other hand, their current chronology does kinda fit the emergence of Derby beer and its lookalikes in the early 1600s. The superior taste that led to Derby beer becoming Britain’s first nationally-known beer “brand” is traditionally ascribed to the use of coke rather than wood in drying the malt, but the dates don’t _quite_ fit. Losing the phenolics because someone had imported a POF- kolsch-type yeast might be a better explanation.

  45. Weehee

    I had a look to see if WLP030 was similar to any of the brewtek yeast strain descriptions but there was nothing of note.

    Can’t see anything on the old user groups. Allegedly the other brewlab thames yeast could be courage.

    There’s some forum threads speculating on the source of 1882 thames valley yeast from wyeast. Maybe they’re the same? https://www.brewuk.co.uk/forums/viewtopic.php?f=8&t=1674

  46. Weehee

    Also it says ESB which is really fuller’s trademark isn’t it – but I guess back then they thought it was a style? Maybe it’s just a variant of their yeast from a bottle.

  47. qq

    The way 1882 is described as from “outside Oxford” sounds less like Henley/Brakspear and more like Abingdon, which means Morrells (incidentally Sharp’s use Morells’ yeast for Doom Bar etc, I assume it is/was a multistrain and 1882 is just part of it).

    Brakspear is complicated, their original multistrain that they used until the 90s was derived from the Mann yeast that got widely used across Southeast England, particularly within the Greene King empire. But then the 90s/noughties turmoil was particularly painful for them and they stopped brewing until they started again in a small way under Marstons’ wing. Kristen/mrmalty appears to have had direct confirmation from Wyeast that 1275 Thames Valley is from Marston via “new” Brakspear which is why he’s put it as an equivalent to WLP023 Burton.

    In the UK ESB is a Fuller’s trademark, in the US it’s a style. The BJCP have invented several styles to accomodate individual British beers (the old Northern & Southern Browns were for Newky Brown and Mann’s respectively even though it was historical nonsense) and since US brewers think that beer begins at 5% ABV, a style of “British but stronger than 90% of traditional British beer” is actually quite popular over there, even if you rarely see such beers in the UK.

    Also the US think of Mackeson as a strong beer by British standards, they got the 4.9% export version.

  48. qq

    Depends what you mean by “if there are/was 5 strains in us-05 I hope it doesn’t compromise any of the studies which used it”?

    If people are using yeast in the lab, they will culture off a single colony as a matter of routine, and people using it in a brewing context are used to the idea of multistrains, so I don’t really see the issue?

    Here’s a 2015 interview with Steven Dresler of Sierra Nevada where he confirms the origin of “Chico” – they took slants from the master “Slant #96” at Siebel in Chicago, but now do all their yeast work inhouse and as it’s taken on its own house character as they’ve taken it through bigger and bigger vessels. And he confirms that as of 2015, they used the production strain for bottling the Pale Ale and don’t pasteurise it (although they do for the Bretty stuff).
    https://www.homebrewtalk.com/brewmaster-interview-steven-dresler-sierra-nevada.html

    Looking at the dendrogram BRY-96 and BRY-97 are close enough that it looks like a single yeast strain came across the Atlantic a century ago, and diverged slightly in the two Ballantine’s brewhouses before being captured by Siebel.

  49. Ed

    For some reason I can’t reply under the relevant post. Defeated by technology again. Anyway, I’m terrible with names but the gentleman from Surebrew is Chris Giles. If I remember rightly he said the differences between the various strains found in US-05 was in flocculation characteristics, which could well mean there’s mutation from the original parent strain rather than it originating as a complex multi-strain mix.

  50. Weehee

    I think my original idea was that if you are using a packet of yeast for a study and there are actually more strains than expected it might not be reproducible but that’s a fair point.

    Although it I guess if there are 5 strains there’s no way of knowing if it’s on purpose or not and if the composition/ratio changes; I’d have thought if you changed the yeast composition of an experiment then it might change the outcomes. I’m curious to what else is in there now…

    That’s pretty cool, thanks for the link. I’m planning isolating some yeast from their pale ale soon.

    I wonder if it was fermenting the strain cold which made it split off. The other different characteristic between them is 97 being a top cropper. 97 is also more flocculent isn’t it?

  51. qq

    You only get reply buttons in the top 2 or 3 levels of a thread, after that you have to go up a few posts to find the reply button.

    Makes sense that what we’re seeing with US-05 is a single strain splitting by flocculation, sounds very similar to Gilliland’s work on the Guinness yeast. But US-05 is marketed as a Chico equivalent, whereas things like Nottingham don’t pretend to be a single strain. When I hear rumours that Mangrove Jack are particularly bad when repitched, I wonder what they’ve been mixed with – and it might explain the recent reformulation of some of their yeasts, they now come with something that discourages reuse?

  52. qq

    Interesting where those Danish wild ones fit, they’re cold-tolerant POF+ at the right-hand end of the Mixed clade, in amongst some bread yeasts, Beer023/5 (Belgian refermentation) and WLP036 Dusseldorf. Judging from that presentation slide, they must be right in there close to the origins of the Saaz lager yeasts as well. Which is interesting.

  53. Weehee

    I sent them an email a while ago about whether or not their lager strains were pastorianus, they are all cerevisiae with one of them being a blend. I’ll dig up the email if you’re interested in which strain they said was a blend.

  54. bierhaus

    Great stuff guys, really interesting. That said, I saw my name come up in the comments about Conan yeast and thought I’d clarify some things. In particular, I am no longer certain Conan is NCYC-1188. What initially led me to believe that came from a MBAA presentation I attended where a former VPB employee who worked at the pub in the early 90’s said (off the record) that their primary ale culture was the non-top cropping portion of the Pugsley Ringwood yeast. At the time, it was believed that NCYC-1187 and 1188 represented the two primary cultures from this yeast blend. We now know that the original Hull-Ringwood culture contained over 4 strains and Pugsley selected two of them for use in his breweries. So even if NCYC-1187 is one part of the Ringwood culture, the number that corresponds to the other primary yeast is uncertain. Complicating matters is that VPB used many different strains during its formative years and repitched some yeasts continuously. I’ve since heard that Conan is actually Scottish yeast
    – although I have a hard time believing that! Lastly, I also find it highly unlikely that Noonan came up with his “VPB-1188” designation out of the blue.

    What I can say is that Conan yeast (as harvested from Heady) is amazingly similar to the non-head forming portion of the Ringwood culture still in use in the North East (Shipyard, Geary’s, ect). I have both in my collection and have done multiple side by side comparisons. Overall, they are pretty darn similar in taste and ester profiles when tested via GCMS. It would be interesting to see some PCR work done on these.

  55. qq

    VPB-1188 could equally be a bit of an inside joke at Austin/Pugsley’s expense, using the fact that VPB opened in November 1988. It’s possible it came from Pugsley – but he came to the US in 1986, which doesn’t quite fit with John Kimmich saying it’s “the yeast that Greg brought back from England, years ago…that was Greg’s discovery”. There could have been contact in the UK, we don’t know. I don’t think we can draw too many assumptions either way – let the DNA do the talking. That’s not to dismiss your GCMS results, that’s interesting – but I’m a DNA man, so that’s the hammer I like to use when I want to crack a nut….

    At the moment the only DNA we seem to have on Conan is isomerization’s
    delta2/12 PCR : http://oi67.tinypic.com/2112yiu.jpg

    The Yeast Bay Conan, WLP4000, is lane #9 with a distinctive band at ~900bp. That band is even stronger in lane #5 (1968 “Fuller’s”), and weaker (not just a loading effect by the look of it) in #2 (1272 BRY-97/WLP051) and #3 (1332 “Hales/Gales”). That doesn’t mean it is the Fuller’s yeast, or even a close cousin it’s just one bit of the genome that looks a bit similar, but it whets the appetite for better data… As an aside, it’s curious that 1272 in lane #2 is so different to the Chicos in lane #1 (1056) and #13 (US-05) whilst being so similar to lane #3, I’d…like to see that repeated.

    One imagines that when Ringwood set up in 1978 they went through a similar process to these guys in 1981 – sounds like they ended up with a nice yeast, I’ve had a quick go at identifying it in the NCYC catalogue but didn’t get very far. I’m tempted to get a bottle, I”m pretty sure there’s yeast to be harvested there.
    http://www.burtonbridgebrewery.co.uk/History/Index.shtml

  56. Trevor

    I’m following anxiously the reveal of Saaz-type vs Frohber-type yeast reveals. I can’t find much information right now on what those are other than this (from some forum comments on AHA):

    Saaz strains: WY2000 (so this must not be equivalent to WLP800?), WY2001, WY2278
    Frohberg strains: WY2124-34/70-WLP830, WLP800 (assume as much based on news it is S. Ceriv.)

    Really curious about WY2782, WY2002, and WLP802 and if all Saaz-type strains can’t use maltotriose or just some.

  57. suregork Post author

    Unfortunately there hasn’t been any progress yet on the Saaz- and Frohberg-type reveal of the commercial strains, so can’t comment on your list of strains.
    Phil (Weehee here) was going to send over some strains, but we haven’t heard from him in months; hopefully he is okay.

    Kristoffer

  58. Trevor

    I would be happy to send over some strains if it would help. I don’t have an issue buying and shipping packs of WY27782, WLP802, and L28 (WLP2002 not for sale currently) if that would make it happen.

  59. suregork Post author

    Thanks for the offer! I’m quite busy at the moment, but I’ll get back to you when I have time to start up this project again 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *