TSS normalization for relative abundance input in MaAsLin 3

Downstream analysis and statistics
1

Hi Will and MaAsLin group,

The MaAsLin 3 tutorial recommends TSS normalization, but its example appears to use data that are already expressed as relative abundances. Why apply TSS again? I’m using relative abundance data generated by MetaPhlAn 4, in case that affects the recommendation. Should I use normalization = "TSS" or normalization = "NONE"?

Thanks!

2

Hi,

If you apply TSS normalization to something that’s already TSS normalized you just get the same thing back (if the abundances already sum to 1, scaling them to sum to 1 doesn’t change anything). Therefore, using the TSS parameter works both for count data that are not yet scaled as well as MetaPhlAn style output that’s already scaled. In your case, I’d just use the default TSS.

Will

1 Like