search

dviraran / SingleR

SingleR: Single-cell RNA-seq cell types Recognition (legacy version)
GNU General Public License v3.0
266 stars 98 forks source link

CreateSinglerObject in Window #53

Open whdmstjr0702 opened 5 years ago

whdmstjr0702 commented 5 years ago

Hello,

I am using SingleR package v0.2.2 in Window.

If I try to run the code: SingleR_Code

Than I get following error message: SingleR_Error

If I do manually assign number of CPU core to 1, I still get following error message: SingleR_Error_CPU_1

Is there no other way to increase number of cores to use in Window?

Also, can you take a look at the error message regarding missing TRUE/FALSE value?

Thank you.

dviraran commented 5 years ago

Oh, wow. You are probably the first one trying to use it on Windows... :) It is possible to dig in the code and change the parallelization to packages supported by windows (now I'm using pbmcapply, which is not supported). Maybe in future versions.

The other error does not seem to be related to the number of cores. I think this happens when none of the references data genes is found in the single-cell data. You can see that it already starts pretty low - only 947 genes are used in the first round (in Immgen this is usually over 2000 genes), so the fine-tuning step has very low number of genes to work with. Why is it so low?

whdmstjr0702 commented 5 years ago

Hello,

Thank you for quick reply.

So I guess, likely I have to run it in Mac for now.

In regards to number of genes, I am trying to figure that out through Seurat GitHub.

Thank you.

Ed

natashiabo commented 5 years ago

I also get this error and the process is very slow:

Capture_scn

Capture_scn2

dviraran commented 5 years ago

Yes, it is pretty slow.

Windows doesn't support the parallelization, so you need to use numCores = 1 if you are on Windows OS.