Whether the data after integration could be used for DE analysis about liger HOT 6 CLOSED

LIGER does not modify the original gene expression matrix during integration. Thus, you can perform standard differential expression analysis to identify cell-type-specific differences between the datasets you're integrating. Alternatively, you can use the dataset-specific metagene loadings (V matrix) directly to identify DE genes. You can see examples of how we did this in Figures 3-4 of the paper.

from liger.

The function getFactorMarkers, and visualization functions plotWordClouds, plotGeneLoadings are designed to help identify DE genes based on the dataset-specific factor loadings. Alternatively, you can apply DE tests of your choice on the raw data after assigning clusters -- converting to a Seurat object to use their DE test suite might help here.

from liger.

Thanks for the advice! I will try them. ( now stuck in the installation step)
> install_github('MacoskoLab/liger') Downloading GitHub repo MacoskoLab/liger@master MacoskoLab-liger-b0b1eae/vignettes/pbmc_alignment.zip: truncated gzip input tar.exe: Error exit delayed from previous errors. Error: Does not appear to be an R package (no DESCRIPTION) In addition: Warning messages: 1: In utils::untar(tarfile, ...) : ‘tar.exe -zxf "C:\Users\zju\AppData\Local\Temp\RtmpAnUFlr\file3f106f1f6124.tar.gz" -C "C:/Users/zju/AppData/Local/Temp/RtmpAnUFlr/remotes3f104814445b"’ returned error code 1 2: In system(cmd, intern = TRUE) : running command 'tar.exe -ztf "C:\Users\zju\AppData\Local\Temp\RtmpAnUFlr\file3f106f1f6124.tar.gz"'
I have checked the java version and the umap tool and I have no idea why the installation failed. There is no problem when downloading other package from github.

from liger.

What versions of R and Rstudio are you using? This may be related to this issue. One solution seems to be to set Sys.setenv("TAR" = "internal") before doing the install.

from liger.

What versions of R and Rstudio are you using? This may be related to this issue. One solution seems to be to set Sys.setenv("TAR" = "internal") before doing the install.

Thanks for your advice. R version is 3.5.2 and Rstudio version is 1.1.463. The system of my computer is windows10. I have tried to use Docker by:
docker run -d -v /mypath/:/home/rstudio -p 8787:8787 docker.io/vkozareva/sc-liger:latest
It works smoothly. Next I could analyze my data. Thank you so much.

from liger.

Great, I'll go ahead and close this then!

from liger.