The consistent monitoring of yeast viability and growth is an important element of fermentation processes. The ASBC hemocytometer count technique is the method most commonly used for performing this.
In this technique, a hemocytometer is used to manually count samples once they have been taken from the fermentation vessel and stained with methylene blue.
While this technique is documented and widely known, at best, it is a prediction that is established on a very small sample count. When observed using a microscope, the hemocytometer displays a grid of measurement areas as seen below.
Figure 1. Image Credit: Yokogawa Fluid Imaging Technologies, Inc.
Due to the time required for the operator to perform manual counting, only a small quantity of actual grid cells are counted, which interpolates the results as an average number.
The sample size is very small, which generates low statistical significance, and it has been established that errors of up to 25% can be added just by the interpretation of the operator.
The goal of this article is to demonstrate a technique using Flow Imaging Microscopy for increasing the precision of yeast counts, eliminating the time and potential for operator error in this process, and increasing the statistical significance by covering a larger sample.
Method
The FlowCam®, a flow imaging microscope, is optimally designed for the automation of this procedure. It can count, image, and quantify thousands of separate yeast cells in the time taken for a user of the hemocytometer technique to count only tens of cells.
A count of dead, budding, and live yeast cells can be automatically performed by the FlowCam's VisualSpreadsheet® software with no requirement for operator involvement. This produces highly repeatable and precise results and normalizes out human error. Due to the larger populations of data acquired by the FlowCam, the results have a much stronger statistical significance.
Similar to the hemocytometer technique, the yeast samples are removed from the fermentation vessel and are prepared by staining them with methylene blue. The sample is then processed by the FlowCam in auto image mode at seven frames per second as it moves through the flow cell. Each individual yeast cell is imaged, saved, and quantified as they are acquired.
Figure 2. Image Credit: Yokogawa Fluid Imaging Technologies, Inc.
As shown in Figure 2, the FlowCam instantly images every yeast cell as a single stored image from the fluid flow. More than 40 morphological properties are indexed to the images of individual cells during the sample run.
Dead cells will uptake the stain when the yeast cells are stained with methylene blue, which results in them appearing blue to the camera. Figure 3 below demonstrates how the cells would be counted in the hemocytometer.
Figure 3. Image Credit: Yokogawa Fluid Imaging Technologies, Inc.
It is simple to differentiate between live and dead cells using the FlowCam by using the ‘average blue’ value captured for the cell image (in combination with multiple shape measurements).
A more complex challenge is presented by the ‘budding’ cells because the resolution required to correctly distinguish a single ‘live’ cell from a ‘budding’ cell is much greater than what can be acquired with the FlowCam.
An easy solution to this challenge is to notice ‘doublets’, which is where two yeast cells have already ‘budded’ and are going to separate. When counting ‘budding’ cells, the operator is mainly looking for signals that the yeast is still growing and viable.
To quantify budding, the system should be filtered to find doublets, and each doublet should be counted as two ‘live’ cells and one ‘budding’ cell. In this case, the trend is the critical measurement rather than the actual number.
Results and Conclusions
Figure 4 below demonstrates how the FlowCam automatically determines the quantity of dead, live, and budding yeast cells. In 35 seconds, the FlowCam automatically characterized a total of 8,709 yeast cells in this experiment.
Figure 4. Image Credit: Yokogawa Fluid Imaging Technologies, Inc.
In contrast with the hemocytometer counts, the FlowCam performs an accurate count of the cells instead of a prediction founded on extrapolation. The concentration for each type of cell is automatically calculated by the FlowCam as part of the system. FlowCam results provide a higher degree of statistical significance due to the large set of data.
As human interpretation errors are eliminated, the FlowCam results offer a high degree of precision over several runs, with a small degree of variability, which is generally as low as 1%.
As outlined before, the filters employed for yeast characterization only need to be defined once. Filters can be re-used for all further samples once they have been defined.
VisualSpreadsheet enables the filters to be easily defined, the user locates particle images of the required type by pressing on them, and then instructs the software to save these as a filter.
Using statistical pattern recognition, the filter then simply locates particles that are similar. The analysis is completely automatic after this stage.
This information has been sourced, reviewed and adapted from materials provided by Yokogawa Fluid Imaging Technologies, Inc.
For more information on this source, please visit Yokogawa Fluid Imaging Technologies, Inc.