Inquiry about the minimum number of input reads


I had a brunch of metagenomic samples pushed through MetaPhlAn 4. However, samples have various number of reads. Some sample have much more reads than others. Even though I have results for each sample, I am not sure if there is a requirement or recommendation for the minimum number of input short reads to get the accurate composition of a microbial community.

Thanks for your help,


This is a quite difficult question and will really depends on the type of sample / environment you want to profile. For they human gut microbiome, we usually set 1 or 2 million reads as the minimum for an accurate profiling

Thank you @aitor.blancomiguez for your reply.

Happy Holidays!