`prep_fcd()`: adjust `max_cell` for each file

[Muffin2001]

Problem

From a set of .fcs files, I need to import a different number of events from each file. prep_fcd() only allows for one max_cell parameter for the whole process. Technically, one could separately generate .fcs files with a specific number of cells for each file e.g. manually from FlowJO, but that's annoying and not open source. Separate prep_fcd() for separate files with subsequent merge_condor() might look like it works, but separate imports mean separate and therefore different transformations, as (according to my understanting) e.g. autologi transformation fits its parameters to the input data.

New prep_fcd_man() function

So I edited the code from cyCONDOR/R/data_load_and_transform.R to make my own prep_fcd_man() function using a vector max_cell_vec instead of max_cell. It's just very few changes in the code, but works very well for me.

For practicability and cleanliness reasons I can recommend this: I keep the code in a separate R script file that I "import" via source(file_path/file_name.R) - alongside other shenanigans, color palettes, and other custom functions.

I would also appreaciate it if it became a feature of the anno table interpretation, so that I could specify in the anno table how many cells to use from which sample, and maybe even a warning message if the desired number of cells is not there.

Source Code

A minimally edited working copy of the edited cyCONDOR/R/data_load_and_transform.R code is attached as .txt file because I don't know how to make pull requests (lol), so you just copy-paste it into an R script, save it and source it from there.
I think I marked all lines of code where I changed it (~10 lines I think), and renamed all changed functions to not impair existing functions. I had to import flowCore for the "read.FCS()" function.

New Input

Here is how the new input looks like in my case:

source(file_path/prep_fcd_man.R)
max_cell_vec_test <- c(50, 40, 30, rep(20, 60), 1, 70000, 20251031)     # here, rep(20, 60) generates 60 vector entries of 20

condor  <- prep_fcd_man(data_path = ("file_path"),
                   max_cell_vec = max_cell_vec_test,           # this is the new input
                   useCSV = FALSE, 
                   anno_table = "anno_table.csv",
                   filename_col = "filename",
                   separator_anno = ";",
                   transformation = "auto_logi",
                   seed = 91)

I would also like to give a shoutout to all the trees that I see through my window

They are loosing their leaves
because it is autumn.

prep_fcd_man.txt

[SamGG]

Hi, excellent suggestion. I've also needed this kind of feature from time to time.
I hope Lorenzo can add a new case directly to the prep_fcd code where max_cell is a vector rather than a single number. I am not in favor of a specific function. We need to check that the length of the vector is the same as the number of FCS files.
Adding a specific column to the annotation table is an excellent alternative, IMHO.
By the way, this could also apply to the seed.

[Muffin2001]

Yeah I had this idea:

prep_fcd(..., max_cell, ...)

if max_cell is number -> use this number for all samples (like current prep_fcd())

elif max_cell is a suitable vector -> use as max_cell_vec like in my prep_fcd_man() function

elif max_cell is column name of anno_table -> extract column and apply like prep_fcd_man(..., max_cel_vec, ...)

else -> max_cell error message

Keeping the simple input option intact for the basic use case, whilst incorporating more functionality for special cases

[lorenzobonaguro]

Yes, I think this is a really good idea! We will implement it in the next release!