Numerical variables cannot be plotted in PCA

[Jakob37]

Seems to be coming from this part of the code, which in my data converted:

> raw_df[["yield_vol"]]
 [1] 0.05 0.09 0.09 0.06 0.09 0.08 0.22 0.13 0.20 0.14 0.08 0.12 0.12 0.19 0.11 0.21 0.12 0.20 0.14 0.11 0.05 0.09 0.10 0.18 0.12 0.16 0.08 0.06 0.05
[30] 0.18 0.26 0.17 0.11 0.15 0.09 0.02 0.15 0.05 0.11 0.19 0.09 0.05 0.16

to

> raw_df[["yield_vol"]]
 [1] "X0.05" "X0.09" "X0.09" "X0.06" "X0.09" "X0.08" "X0.22" "X0.13" "X0.2"  "X0.14" "X0.08" "X0.12" "X0.12" "X0.19" "X0.11" "X0.21" "X0.12" "X0.2" 
[19] "X0.14" "X0.11" "X0.05" "X0.09" "X0.1"  "X0.18" "X0.12" "X0.16" "X0.08" "X0.06" "X0.05" "X0.18" "X0.26" "X0.17" "X0.11" "X0.15" "X0.09" "X0.02"
[37] "X0.15" "X0.05" "X0.11" "X0.19" "X0.09" "X0.05" "X0.16"

OmicLoupe/R/reactive_values.R

Line 21 in 0f41a08

mutate_all(make.names) # Run make.names on all variables of the design matrix

After removing that line I get the desired PCA chart (before it will be a categorical coloring, one value per category):

[Jakob37]

The consequece is that all variables (including numbers) in the design matrix are converted to valid names. I suspect this has been added to solve some other issue, do you know @manszamore ? I would rather revert it and solve that other issue some other way.

I can make a PR for fixing this.

[manszamore]

Yes, I think this was added to allow for using variables (e.g. column names in the design matrix) beginning with numbers. But it's not ideal if it breaks the PCA, of course. Feel free to revert it! :)