Installing Cell Ranger

Question

[gdfsnkfns]

Hello everyone.

I'm now trying to install Cell Ranger following the tutorial described in 10x genomics website (https://www.10xgenomics.com/jp/support/software/cell-ranger/latest/tutorials/cr-tutorial-in), but facing the problem.

I installed the Cell ranger package (cellranger-8.0.0) and added a path to it as follows;

export PATH=/Users/******/desktop/cellranger_downloads/cellranger-8.0.0:$PATH

I suppose Cell Ranger was successfully added to the $PATH because output of "which cellranger" command seemed proper;

/Users/******/desktop/cellranger_downloads/cellranger-8.0.0/cellranger

However, when I runned "cellranger" command or tried testrun (cellranger testrun --id=tiny), the following error message appeared...

**zsh: exec format error: cellranger**

I'm conducting the experiment using Mac PC (Apple M1 Max).

If anyone knows of a solution, I would appreciate it if you could enlighten me.

Best regards,

[Haci]

You cannot run cellranger on a Mac: https://kb.10xgenomics.com/hc/en-us/articles/115003535986-Can-I-run-Cell-Ranger-on-my-Mac-or-Windows-machine

If you do not have access to a GNU/Linux system, the alternatives that I can think of is using docker directly or switching to nf-core/scrnaseq(https://nf-co.re/scrnaseq/2.5.1) and making use of docker/singularity indirectly.

[gdfsnkfns]

Thank you all for all the advice. I was completely mistaken that the Cell Ranger could be operated in a terminal. I will try the analysis using Docker. If I have any problems, I might have to post question here. Thank you in advance.

Answer 1

[GenoMax]

cellranger you downloaded is compiled for Linux. You can't use it as is on macOS. Since celleanger is a closed source application you can't do much about this (i.e. you can't recompile etc).

Consider switching to alevin-fry or starsolo if you do not have access to linux hardware but need to work with single-cell data.

[ATpoint]

If macOS is the only choice I would consider running it through Docker. SInce CellRanger itself is fully self-contained you (probably, unested) don't even need to build a container, but just use any Linux container (CentOS, Ubuntu) and run commands from inside there.

[GenoMax]

Do you know if this will scale for a large number of samples? OP seems to have a M1 Max so may have adequate amount of RAM but one never knows.

[dsull]

"Running Cell Ranger requires at least 8 CPUs, preferably 16, and at least 64 GB of RAM, preferably 128." - https://www.10xgenomics.com/support/software/cell-ranger/latest/tutorials/cr-tutorial-in

Yeah... meanwhile, all other open source programs take a quarter (or less) of that to run.

Cell Ranger still has useful utilities, like bamtofastq, and so it's still worth installing for certain use cases.

[ATpoint]

My personal opinion is that scaling / big data and use of laptops are mutually exclusive or at least not a native choice. You will probably need to allocate most of the resources to the Docker VM in such a scenario. On top, be sure to disable all these (useless) additional analysis that CellRanger does (like tSNEs etc, I think flag is something like --no-secondary) to save time and RAM. If you're at a company or university reach out to IT and ask whether any cluste/HPC infrastructures is available that you could use. If not, then it's just wait and see. A "normal" 10x run (say 10k cells, 30k reads/cell) takes some hours on a decent HPC node, so depending on your number of samples that will take time on a laptop.