how to gauge yeast viability in a microscope

[iandow]

Is it possible to look at a sample of active yeast in a microscope to know if a yeast starter needs another feeding, e.g. because it sat too long or because the first feeding didn't produce enough cells?

I took a picture of some active homebrew yeast with my daughter's microscope today, at 200x magnification. Below is a picture of what I could see. Is 200x magnification enough to count the cells? Do I need to just count cells, or do I need to count multiplying cells (i.e. 2 cells attached together). It would be pretty challenging to count the cells at 200x because there are so many of them. I think professional brewers must use higher magnification.

 

[Vale71]

Actually 100x would be more than enough. If you seriously want to engage in cell counting and viability assessment you'll need a hemocytometer and some methylene blue. It can be lots of fun but it's a bit more challenging than just looking at cells in a microscope, especially if you spill the ethylene blue and ruin your wife's kitchen...

 

[mongoose33]

I attended a Yeast workshop put on by White Labs at the BREW boot camp in March 2019; one day was all the theory; the second day actual work in one of their labs which was very, very cool.

To figure out the number of cells in a starter (or a batch of brewing beer for that matter), you have to take a good sample and then dilute that sample several times using distilled water. You use specific tubes to do that (sort of like test tubes but plastic). I haven't done it in a while, but you sample a CC of beer or starter, add to the tube, add IIRC 9 CC of water, shake it up to mix thoroughly, sample that, and you do it three or four times (I don't recall exactly--don't have my notes handy).

Then you put the resulting solution under a microscope. We were told to use one that does 400x magnification, though if yours works with a hemocytometer, well, then it works. The cells sure look similar in size in your pic as they are in mine.

Then, you pick different squares in the hemocytometer and count yeast cells within those squares which is one heck of an easier way to do it. And, anyone who does this professionally uses a counter/clicker instead of just saying "one, two, three...."

In a 5x5 grid of squares, for instance, you'd count the yeast in the four corner squares and the middle one, average the result, then use a multiplier to determine/estimate the actual number of cells, then multiply it back to determine the total number.

And as Vale71 says, you use ethylene blue to stain the cells. The live cells can keep the stain out; dead ones turn blue.

Below are two pics I took to show this. The first shows some yeast including a few that have turned blue. The second shows how the hemocytometer looks under the microscope. You can see the cells look like yours, except there are mostly blue ones. Kara, who taught the lab part of the workshop, nuked a sample in a microwave so we'd see what a mostly dead culture of yeast would look like. Most of the cells in the second pic are dead, but there are still a few hanging on.

Here are a couple more pics. The first shows yours truly peering into a microscope at White Labs. The second shows the kind of plastic tubes we're using.

Below are links to the tubes, microscope I have (which is what White Labs had ), clicker, and hemocytometer.

https://www.amazon.com/gp/product/B00UJGNU1G/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
https://www.amazon.com/gp/product/B0094JTZOU/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
https://www.amazon.com/gp/product/B071W3QTBX/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
https://www.amazon.com/gp/product/B076ZT949V/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1

 

[mongoose33]

One more thing. Perhaps the above is more than you wanted to know. I went to that workshop because I believed at the time that the next leap forward I could make in my brewing was to understand and manage yeast better.

Well, maybe. I'm retiring literally at the end of next week, so I'll have a lot more time to play with this, but what I realized is that whatever yeast management practices I was using, and my methods to do starters and initiate fermentations, were working extremely well.

That is, I was (and still am) focused on avoiding off-flavors and producing flavor that I wanted to have in my beers. Guess what? No apparent off flavors--and I have a couple people also taste-testing on that for me--and I have friends who want to buy my beer at commercial prices, and a local bar that wants to sell it. So I think it's pretty decent, plus *I* like it. Which is the whole point.

I will spend more time with this once I'm retired, but clearly what I'm doing is working. At the workshop mentioned above Chris White noted essentially that at the homebrew level, the process is pretty resilient. In fact, he said he'd likely just pitch tubes of yeast without even doing a starter. I've tried that a couple times and concluded he's probably right.

At the commercial level, it's a different deal. At the homebrew level? Probably not so much. So if you're doing a good starter, unless it's a really big beer, you're likely doing just fine, and counting yeast and so on isn't, IMO, likely to reward you all that much.

And in the end, the only criterion that really matters is....how does the beer taste?

Now, I'm not trying to dissuade you from pursuing this further--and perhaps, when I spend more time with it, I find something that really elevates my beer to a higher level. Or I learn other things that cement information better in my mind. And certainly, what the hell? Why not learn as much as we can about everything we can?

 

[Vale71]

And, anyone who does this professionally uses a counter/clicker instead of just saying "one, two, three...."

Actually, anyone who does this professionally uses very expensive automated equipment like this one.

https://www.nexcelom.com/nexcelom-p...bS8rQBI1AJBepR8sY1-GXyL7ffCn0COMaAh66EALw_wcB
Manual counting is really way too labour intensive and besides that it's also wildly inaccurate. Test the same sample ten times and you'll see how wildly the result can very between iterations. The viability assessment with methylene blue is also extremely unreliable: dead cells can fail to stain and appear to be alive and kicking and live cells can fail to expel the die quickly enough and end up being counted as dead.

But it's sure fun to do and you learn a lot but as far as quality control is concerned anybody who doesn't work at Coors or Pabst scale is probably better served just following best practices.

 

[mongoose33]

Actually, anyone who does this professionally uses very expensive automated equipment like this one.

https://www.nexcelom.com/nexcelom-p...bS8rQBI1AJBepR8sY1-GXyL7ffCn0COMaAh66EALw_wcB
Manual counting is really way too labour intensive and besides that it's also wildly inaccurate. Test the same sample ten times and you'll see how wildly the result can very between iterations. The viability assessment with methylene blue is also extremely unreliable: dead cells can fail to stain and appear to be alive and kicking and live cells can fail to expel the die quickly enough and end up being counted as dead.

But it's sure fun to do and you learn a lot but as far as quality control is concerned anybody who doesn't work at Coors or Pabst scale is probably better served just following best practices.

Well, that's not really true. I don't doubt that there are some places doing it as you note above, especially the majors, but I visited a brewer a little over a year ago, sat with their guy whose job it was to....count yeast.

So when you say "Actually, anyone who does this professionally uses very expensive automated equipment like this one" it may be true that some do it, but that blanket statement is untrue.

That's why we need to be careful in making statements in the absolute; unless we know it to be true without exception, then the statement is simply one of deception.

And we don't want to deceive people, do we?

 

[Vale71]

Admittedly there are some nostalgic types that probably still do it. I'd still argue that they're just wasting their time and money because of the high error margin but if they're happy with it then who am I to argue with them?

 

[isomerization]

Actually, anyone who does this professionally uses very expensive automated equipment like this one.

https://www.nexcelom.com/nexcelom-p...bS8rQBI1AJBepR8sY1-GXyL7ffCn0COMaAh66EALw_wcB
Manual counting is really way too labour intensive and besides that it's also wildly inaccurate. Test the same sample ten times and you'll see how wildly the result can very between iterations. The viability assessment with methylene blue is also extremely unreliable: dead cells can fail to stain and appear to be alive and kicking and live cells can fail to expel the die quickly enough and end up being counted as dead.

But it's sure fun to do and you learn a lot but as far as quality control is concerned anybody who doesn't work at Coors or Pabst scale is probably better served just following best practices.

Homebrewers without pipettes might be wildly inaccurate, but our manual counts are within 10% accuracy of the automated counts (we use a Vi-Cell XR). I work in a GMP environment, so these things are definitely important. With that said, manual counting is avoided at all costs when automated is available!

I would be interested to know just how accurate you need to be with pitch rate for repeatability between the same batch.

 

[Vale71]

Homebrewers without pipettes might be wildly inaccurate, but our manual counts are within 10% accuracy of the automated counts (we use a Vi-Cell XR). I work in a GMP environment, so these things are definitely important. With that said, manual counting is avoided at all costs when automated is available!

I would be interested to know just how accurate you need to be with pitch rate for repeatability between the same batch.

My claim that chamber-based cell counting (with optional staining) is not used any more was probably a slightly exaggerated generalization. As Mongoose gracefully pointed out this is just a part of my plot to take over the world by spreading false and deceitful information via Internet forums...

Your post on the other hand is a good example of how people often do not fully understand what sources of error we have to contend with. Your assessment of 10% error based on comparison between manual and automated counting is misleading and probably grossly underestimated. Before you even get to the actual counting there is sample collection and preparation and those have their own errors that must be taken into account. I don't know about your working environment but in the context of a brewery unless you're doing cytometry on the finished product at some processing stage (f.e. pre- and post-centrifugation) you'll be dealing with yeast stock in the form of slurry. Slurry can be quite disomogeneous (because of the yeast's flocculation characteristic, among other factors) so that sample collection alone can already be the source of a significant error. You then have the necessary dilution steps to perform as counting chambers have a limited range of density outside of which counting becomes either wildly inaccurate because of the small number of cells counted or impractical because clumping would make it impossible to reliably distinguish individual cells. When measuring slurry the dilution is around 1:1000, if measuring an agitated starter it can be just 1:100. This step introduces its own error. Let's say the error is 10% for each zero after the 1. For a 1:1000 dilution it could be as high as 30%, for 1:100 it could be 20%.

Let's assume we are measuring yeast stock that is in a form of a slurry. Let's say we have a 10% error at sampling because of dishomogeneity, 30% because of dilution and a further 10% because of manual counting. The total margin of error is actually 50%. You can verify this by repeating the process from the start and ideally having different individuals perform the process, you don't verify it by feeding the same sample to an automated counter as that only measures the operator induced error in the counting step. If you did that I bet you'd be surprised at the result. 30 to 50% error is really what you'll probably end up with in this context.

Now would you, as a homebrewer, be happy with 50% accuracy in your density measurements? I strongly doubt that and you'd be even less happy about that in a commercial setting, where errors mean money. Hence my comment that as a smaller operation that can't afford very expensive equipment you're much better off just rigorously following best practices rather than relying on wildly inaccurate measurements.

 

[isomerization]

My claim that chamber-based cell counting (with optional staining) is not used any more was probably a slightly exaggerated generalization. As Mongoose gracefully pointed out this is just a part of my plot to take over the world by spreading false and deceitful information via Internet forums...

Your post on the other hand is a good example of how people often do not fully understand what sources of error we have to contend with. Your assessment of 10% error based on comparison between manual and automated counting is misleading and probably grossly underestimated. Before you even get to the actual counting there is sample collection and preparation and those have their own errors that must be taken into account. I don't know about your working environment but in the context of a brewery unless you're doing cytometry on the finished product at some processing stage (f.e. pre- and post-centrifugation) you'll be dealing with yeast stock in the form of slurry. Slurry can be quite disomogeneous (because of the yeast's flocculation characteristic, among other factors) so that sample collection alone can already be the source of a significant error. You then have the necessary dilution steps to perform as counting chambers have a limited range of density outside of which counting becomes either wildly inaccurate because of the small number of cells counted or impractical because clumping would make it impossible to reliably distinguish individual cells. When measuring slurry the dilution is around 1:1000, if measuring an agitated starter it can be just 1:100. This step introduces its own error. Let's say the error is 10% for each zero after the 1. For a 1:1000 dilution it could be as high as 30%, for 1:100 it could be 20%.

Let's assume we are measuring yeast stock that is in a form of a slurry. Let's say we have a 10% error at sampling because of dishomogeneity, 30% because of dilution and a further 10% because of manual counting. The total margin of error is actually 50%. You can verify this by repeating the process from the start and ideally having different individuals perform the process, you don't verify it by feeding the same sample to an automated counter as that only measures the operator induced error in the counting step. If you did that I bet you'd be surprised at the result. 30 to 50% error is really what you'll probably end up with in this context.

Now would you, as a homebrewer, be happy with 50% accuracy in your density measurements? I strongly doubt that and you'd be even less happy about that in a commercial setting, where errors mean money. Hence my comment that as a smaller operation that can't afford very expensive equipment you're much better off just rigorously following best practices rather than relying on wildly inaccurate measurements.

I should have specified that I am working with mammalian cells (3-5x larger) not yeast cells. I haven’t tried counting yeast.

If you’re that concerned about error propagating through sampling, take multiple initial samples and average. Some of that error uncertainty will be eliminated. This type of error will occur with manual or automated, so doesn’t really address my rebuttal of your manual versus auto comments.