search

mikemc / speedyseq

Speedy versions of phyloseq functions
https://mikemc.github.io/speedyseq/
Other
45 stars 6 forks source link

Compare the output of different psmelt versions #5

Closed mikemc closed 5 years ago

mikemc commented 5 years ago

and edit psmelt to remove the differences, or document the differences where speedyseq's output seems preferable and shouldn't break phyloseq function. (e.g., the rownames in phyloseq's psmelt output don't seem to have any meaning and would probably be better left out). Note, the differences seem to be just

  1. row names - speedyseq's psmelt doesn't use them; phyloseq's psmelt does but I'm not sure of the meaning and think they can be dropped
  2. row order differs
  3. phyloseq drops columns from the tax_table (and perhaps sample_data) that are all NA; speedyseq keep's them
mikemc commented 5 years ago

The difference in row order is due to ties in the Abundance. Both versions sort rows by Abundance in decreasing order, but how rows are sorted with Abundance ties depends on the order of the rows prior to the abundance sorting. It seems fine to treat the row order w/in Abundance ties as arbitrary.

mikemc commented 5 years ago

Addressed by cddbc0a5aeeaaf976475b8fec3fb3f8bee40e301