Why Didn’t My Manual Cell Counting Always Work?

Ten years ago, I did my PhD on mouse embryonic stem cell (mESC) differentiation towards pancreatic progenitors. We discovered that seeding cells at a very low density (2,000 cells per cm2) was beneficial; the cells simply responded better to any medium we concocted.

Growing two to five cell lines at a time, passaging cells every two days for five years, I spent a lot of time counting cells in a hemocytometer. When I left academia, I thought I would never look back at cell counting again!

Talking to Cell Counting ‘Nerds’ Taught Me Something: It Does Matter!

Two years ago, I started working at ChemoMetec, a company all about cell counting. Talking with my colleagues in R&D and Sales Support, I realized that manual cell counting holds a lot of bias:

Many of us do not rigorously work to eliminate these sources of error every time we set up a new experiment. But does it even matter?

Well, consider having 10% variation from each of four factors mentioned above. This could give you a combined variation of 40% from experiment to experiment, and you would find yourself getting conflicting results in many experiments. On the plus side, you’d occasionally identify hugely robust protocols that give consistent results. But most often, this would not be the case. In contrast, having a reliable automated cell counting system would provide a very low variation from experiment to experiment, and more experiments giving you significant data.

The Ten-Year Blues Hit Me Hard

This issue does not only go for academic work in petri dishes. Imagine setting up a stirred-tank culture of cells producing recombinant proteins or antibodies, and having a protocol with an optimal harvest timepoint e.g. on day eight when you reach 10x cells per ml. If your cell counting determinations are off by 10% due to a combination of factors, you might need to harvest 10-20 hours sooner or later.

Manual cell counting gives high variation, especially if several people are working together on a project, as their errors in counting and pipetting add up. Many automated cell counters will even give you a combined variation of ≥ 20%, which is detrimental to the robustness of your protocols/SOPs and your efforts to set up a reliable production system.

So, was my PhD work in vain? Well, many of my results did point in the same direction and thereby provided reliable answers. But my advice: Consider your cell counting protocol and adjust if/when beneficial to you. You might gain from it, and you could even avoid the cell counting blues down the road…

Further Reading


  1. K Takahashi and S Yamanaka: Induction of Pluripotent Stem Cells from Mouse Embryonic and Adult Fibroblast Cultures by Defined Factors. Cell. 2006; 126, 663–676.
  2. K Takahashi, K Tanabe, M Ohnuki et al.: Induction of Pluripotent Stem Cells from Adult Human Fibroblasts by Defined Factors. Cell. 2007; 131, 861–872.
By Janny Marie L Peterslund, Scientific Affairs Manager at ChemoMetec

Janny Marie has a background in stem cell research at Novo Nordisk, Denmark. She worked at STEMCELL Technologies, supporting and managing the epithelial cell culture product portfolio. Now at ChemoMetec, Janny Marie supports collaborations with external partners and writes for the Cell Counting Blog.