error message during installation of too-many-cells on ubuntu

[zmokhtari]

I am trying to install too-many-cells on Ubuntu 19.4. I am facing this error message: "-- While building package glib-0.13.6.0 using: /tmp/stack24744/glib-0.13.6.0/.stack-work/dist/x86_64-linux-tinfo6/Cabal-2.2.0.1/setup/setup --builddir=.stack-work/dist/x86_64-linux-tinfo6/Cabal-2.2.0.1 build --ghc-options " -fdiagnostics-color=always" Process exited with code: ExitFailure 1 Progress 211/248" Could you please help me to solve the issue? Thanks in advance, Zeinab

[GregorySchwartz]

This issue is solved with newer glib releases. I updated glib here, so pull the new version and try to install again (and let me know if it works).

[zmokhtari]

thanks for your fast response. Installation worked fine, however, when I try the example in R I get this error : "sh: 1: too-many-cells: not found Error in system2("too-many-cells", args = c(args, autoArgs), stdout = TRUE) : error in running command In addition: Warning message: In dir.create(output, recursive = TRUE) : 'out' already exists" I should mention that I installed toomanycellsR first and then I installed too-many-cells.

[GregorySchwartz]

Docker or from source? If using docker, be sure to set the docker argument.

[zmokhtari]

I have both, but I want to use the source.

[GregorySchwartz]

Is too-many-cells in your path? It should be in ~/.local/bin.

[zmokhtari]

Yes, it is: Copied executables to /home/zeinab/.local/bin:

• too-many-cells

[GregorySchwartz]

Is /home/zeinab/.local/bin in your $PATH?

[zmokhtari]

Yes, I double checked, it is in my $PATH list.

[GregorySchwartz]

Ah, so which too-many-cells is returning a result? I keep bringing this up because sh: 1: too-many-cells: not found means that your shell cannot find the command. If which too-many-cells does return a result, then there must be some disconnect between R and your shell.

[zmokhtari]

It returns"/home/zeinab/.local/bin/too-many-cells". Could you please help me to check the connection between R and the shell?

[GregorySchwartz]

Interesting. Can you open R in your home folder and run system2("too-many-cells", stdout = TRUE)? Then open R in your ~/.local/bin folder and do the same?

[zmokhtari]

thank you for your help. both of them return: "Missing: (make-tree | interactive | differential | diversity | paths)

Usage: too-many-cells (make-tree | interactive | differential | diversity | paths) character(0) attr(,"status") [1] 1 Warning message: In system2("too-many-cells", stdout = TRUE) : running command ''too-many-cells'' had status 1 "

[GregorySchwartz]

So it can find the command in both cases, but in the tooManyCell() function in those same sessions it does not find the command?

[zmokhtari]

That is right! Still I could not solve the issue.

[GregorySchwartz]

This may be beyond my understanding. The tooManyCells() function uses system2("too-many-cells") with no changing of directories, so I don't know what to do if system2 can find the command line program outside the function but not inside (which should not be happening). My suggestion would be to just call it outside the function or use the command line program.

[zmokhtari]

Thanks for your suggestions. I used the command line for the example data and it worked fine, I am wondering if I provide cytof data in csv format, whether too many cells is accepting that format? Do I need to provide gene names and cell IDs too? Can I send you an example of the csv format that you could examine the format of the file? Thanks in advance.

[GregorySchwartz]

You can use csv format, just input it with -m data.csv. The format is a normal csv file, with a header for the name of the column (cell) and an initial column containing the features as the first column. With this data, don't forget to change the default normalization and filters (I would recommend something like --normalization UQNorm and -F for no filtering)!

[zmokhtari]

Thank you very much for your help. It works fine.

[zmokhtari]

Hi Gregory, I am wondering if I provide gene.csv file for the list of features that is measured, how can I point to a specific gene when I want to color the dendrogram based on a gene expression level? thanks

[GregorySchwartz]

The instructions would be the same as on the website (https://gregoryschwartz.github.io/too-many-cells/) under "gene expression". Although you wouldn't have a "genes" file, you just load in the matrix you used earlier to create the tree.

[zmokhtari]

Thanks for your help.

[GregorySchwartz]

My pleasure!