385 followers
@norabelrose Interesting! It seems something similar has been observed in transcriptomics count data, where the count distribution morphs from Zipf to exponential decay, and it’s believed to indeed be a finite sample issue. https://t.co/AxlzSnmceF