tinyi
commented
2 years ago Thank you for your interest in our work. Here are the answers to your questions:
1.
For my single cell data only cell types information is available. I am not sure how to subtype these cells to come up with the cell states. For now I am using the cell types as both cell.type.labels and cell.state.labels as it is mandatory to provide both. Let me know if it is ok.
Yes. If you do not have cell states, you can certainly define them using the cell state label.
I noticed that you have used the argument key="NULL", which should be key=NULL. I guess it was a typo, as otherwise new.prism will throw an error.
2.
The initial data preparation went very well for my data. I just converted the single cell data sparse matrix to dense matrix and I selected only protein coding genes for Prism construction.
It looks like your memory is exhausted. Try remove all variables other than myPrism from your R environment and then clean up the memory using gc(). You may also try reducing the number of threads.
Best,
Tinyi
On Tue, Jun 28, 2022 at 1:50 AM monalisa6hota @.***> wrote:
Hello! I am trying to run BayesPrism for my bulk RNAseq data sample size ~200 and Single cell data, no of cells ~70000. I am working with non-malignant samples. Here is the issue I am facing. Please let me know how to solve this.
1.
For my single cell data only cell types information is available. I am not sure how to subtype these cells to come up with the cell states. For now I am using the cell types as both cell.type.labels and cell.state.labels as it is mandatory to provide both. Let me know if it is ok. 2.
The initial data preparation went very well for my data. I just converted the single cell data sparse matrix to dense matrix and I selected only protein coding genes for Prism construction.
My code is here: myPrism <- new.prism( reference=sc.dat.filtered.pc, mixture=bk.dat, input.type="count.matrix", cell.type.labels = cell.type.labels, cell.state.labels = cell.type.labels, key="NULL", outlier.cut=0.01, outlier.fraction=0.1, ) I am running this job in a cluster with 120GB memory. The job runs for few minutes, provides the cell state information and then terminates with this error
I am getting following error
> number of cells in each cell state
> cell.state.labels
> PJ017-tumor-6 PJ032-tumor-5 myeloid_8 PJ032-tumor-4
> 22 41 49
/sw/Containers/singularity/bin/run_singularity: line 28: 42360 Killed singularity $*
— Reply to this email directly, view it on GitHub https://github.com/Danko-Lab/BayesPrism/issues/3, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB4NHSYCMSFTA6MNFPPSGQTVRKHBBANCNFSM52A35BTQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>
MonalisaHota
Hello! I am trying to run BayesPrism for my bulk RNAseq data sample size ~200 and Single cell data, no of cells ~70000. I am working with non-malignant samples. Here is the issue I am facing. Please let me know how to solve this.
For my single cell data only cell types information is available. I am not sure how to subtype these cells to come up with the cell states. For now I am using the cell types as both cell.type.labels and cell.state.labels as it is mandatory to provide both. Let me know if it is ok.
The initial data preparation went very well for my data. I just converted the single cell data sparse matrix to dense matrix and I selected only protein coding genes for Prism construction.
My code is here: myPrism <- new.prism( reference=sc.dat.filtered.pc, mixture=bk.dat, input.type="count.matrix", cell.type.labels = cell.type.labels, cell.state.labels = cell.type.labels, key="NULL", outlier.cut=0.01, outlier.fraction=0.1, )
I am running this job in a cluster with 120GB memory. This job runs for few minutes, provides the cell state information and terminates with this error.
I am getting following error
> number of cells in each cell state
> cell.state.labels
> PJ017-tumor-6 PJ032-tumor-5 myeloid_8 PJ032-tumor-4
> 22 41 49
/sw/Containers/singularity/bin/run_singularity: line 28: 42360 Killed singularity $*