nFeatures threshold

Question

[fluentin44]

Hi all,

Im QC'ing a dataset and when plotted with an lm line, the cells begin to deviate away from the trendline with increasing features and counts - specifically higher number count number per gene. Do these cells need to be removed from the dataset (i.e. a threshold of ~7500) or are these cells which just happen to have been sequenced more deeply?

Thanks, Matt