Kegg rare removal cut-off

[lonsdzoe]

Hi,

I've used the shogun (burst) pipeline to process a batch of shallow shotgun data. Would you recommend removing rare kegg pathways/modules before conducting statistical analyses? If so, is there a standard cut-off point that should be used?

Thanks,

Z