Heatmap normalise_by only works by sample names

[bernt-matthias]

According to the docs I can

A variable or a specific sample in the metadata to normalise the counts by.

If I do

amp_heatmap(AalborgWWTPs, normalise=T, normalise_by = 'Plant')

I get: Error: Plant is not found among group names, cannot normalise

I debugged this a bit, and it seems that normalisation is only possible by a sample at the moment.

Note, For me it would be fine to fix the docs.

[KasperSkytte]

Hi. It uses the values in the variable supplied to group_by. So fx group by Plant, then supply normalise_by = "Aalborg West" should work. Could adjust docs if it's unclear.

[bernt-matthias]

Just tried

> data("AalborgWWTPs")
> amp_heatmap(AalborgWWTPs, normalise=T, normalise_by = "Aalborg West")
Error: Aalborg West is not found among group names, cannot normalise

[KasperSkytte]

Like this, but a different error. It's rarely used, but should be a quick fix.

library("ampvis2")
#> Loading required package: ggplot2
amp_heatmap(
  AalborgWWTPs,
  group_by = "Plant",
  tax_aggregate = "Genus",
  normalise_by = "Aalborg West"
)
#> Error in normalise_by %chin% unique(abund7$.Group): table is type 'integer' (must be 'character' or NULL)

^{Created on 2024-01-23 with reprex v2.1.0}

[bernt-matthias]

I see. Combining group_by and normalise_by seems to make sense. At least I would have an intuition what happens in the normalisation.