Closed RubyLiu206 closed 6 hours ago
Hi,
Thanks for your support, the same issue could be found in #70
Hi,
Thanks for your support, the same issue could be found in #70
Thanks for forwarding. I actually successfully installed under HPC. However, I have a follow-up question, do you think VAE over-integrating, or over-performed that make all samples with the same cell type composition. Refer as the below image:
I used our in-house bulk rna data, and GEO immune cell data (n_obs × n_vars = 93900 × 28638), using 2000 genes for cell type deconvolution.
Hi @RubyLiu206 ,
I don't quite understand why you chose 2,000 highly variable genes for deconvolution, my tutorial uses raw counts and uses all genes(https://omicverse.readthedocs.io/en/latest/Tutorials-bulk2single/t_bulk2single/), you can scale down the number of cells and don't need to use the 90,000 single cell data as a reference, which speeds up the run.
I really appreciate you quick response!! I see you point. I choose 2k was because of CIBERSORT used signature matrix from 500 to 2000 number of genes. I will try with 28638 genes with 500 as top marker for sure.
Hi,
Thank you for developing this wonderful tool! I was trying to install under my local environment, and I encountered with the following error:
note: This error originates from a subprocess, and is likely not a problem with pip. error: subprocess-exited-with-error
× Getting requirements to build wheel did not run successfully. │ exit code: 1 ╰─> See above for output.
note: This error originates from a subprocess, and is likely not a problem with pip.
I used pip to install, everything else went smoothly except the pip install -U omicverse