Something to think about when doing experiments

I hadn’t thought about this so much before, but even when you’re working with genetically identical cells (i.e. from the same colony), chaos can play a big role. For example, you can have a pretty huge amount of variation in the expression of a specific protein that you’re looking at in isogenic cells (especially when you’re dealing with signalling, where a couple of extra/fewer molecules make a big difference).

Since chaos is a deterministic process, you can control for it by making sure you have the exact same starting point for all cells (this is something we do in yeast already by synchronizing cell cycles prior to counting them). Another thing that I learned of today that helps reduce noise in your data is the following: say you’re using GFP as a reporter – instead of counting just the number of cells that are expressing GFP, you could use a second reporter (like mCherry) attached to a gene that you know is continually expressed under your particular conditions, and normalize the GFP numbers based on what you see with mCherry.

Pretty simple stuff but still worth thinking about when you’re looking at a bunch of single cells and wondering where all the variation might be coming from. Any other ideas about sources of noise in experiments/how to control for them?

 

Leave a Reply