search

ebecht / infinityFlow

26 stars 8 forks source link

Minimum Input cell number #31

Open GalliES opened 5 months ago

GalliES commented 5 months ago

Hello Etienne and thanks for the great package! I have a very simple but crucial question: I am working with human PBMC samples, and it is sometimes tedious to get gazillions of cells. In your experience, is there a minimum amount of input cells required to expect a decent prediction on the LegendPlex? Do you have any experience working with, for instance 8-10k cells/well and 20 parameters backbone panel?

Thanks in advance for your feedback!

Edoardo

ebecht commented 5 months ago

Hi Eduardo, Actually for the models it's OK to have few cells, we studied this in the Supplementary Figure 3 of our paper. But you may run into issues with the staining itself ! For this I have no experience at all, although I would like to test it properly. You can maybe ask BioLegend or Mark Headley (the last author of the same paper) ? If you find information on that part I'm interested in hearing what you have found ! So short answer is computationally it should not be a problem, experimentally it might be but I do not know ! Etienne