How to decide if a cell is of poor quality in single cell data analysis?

Question

[Firingam]

Hi I'm dealing with some data that cane be found here. This dataset is composed of about 5449. I want to eliminate poor quality cells and In order to do this I'm relying on mitochondrial RNA, library size ecc. The question is : Given that I have not so many rows what is a good proportion of counts to retain a cell as an high quality one?

I was thinking maybe 500 reads over 5000 but I'm not sure given the low experience in this bioinformatic field

Thank You in Advance