Please link to the GitHub Discussion for this proposed analysis.
630
Describe the goals of this analysis module.
The goal of this analysis is to perform cell type annotation for non-ETP ALL samples in the portal.
Here are the steps taken:
00. Pre-processing the provided SCE objects
Formatting SCE objects to Seurat objects for downstream analysis
Setting the filter of percent.mt < 25
Features processing: (1) dropping features with 0 counts across all cells and (2) selecting feature that is curated and has the highest count for those features that have multiple gene_IDs mapping to single gene_symbols. For multi-mapped features, features with gene_IDs starting from LINC, SNO, U*, and Y_RNA are dropped. These steps lead to a (3) unique mapping of gene_ID to gene_symbols.
Doublet removal (using annotation from ALSF)
01. Feature selection/dimensionality reduction
Using self-assembling manifold algorithm, which is a soft feature selection strategy for better separation of more homogenous populations
02. Cell type annotation with marker genes
Using ScType to annotate cell type with markers from Azimuth reference (Human - Bone Marrow) and cancer marker from immune system in ScType database
03. Tumor cell identification
Using CopyKat to identify genome-wide aneuploidy in single cell for distinguishing tumor cells from normal cells, by providing annotated B cells as normal cells
What software will you require?
R (4.2.3)
Main packages: Seurat, reticulate,SAM,ScType,CopyKat
Adding analysis code for 00. Pre-processing the provided SCE objects
What computational resources will you require?
I am currently using our own computing system for the analysis. I may need help in transferring files or large objects, if there is any issue encountered in the future.
If known, when do you expect to file the first pull request?
Please link to the GitHub Discussion for this proposed analysis.
630
Describe the goals of this analysis module.
The goal of this analysis is to perform cell type annotation for non-ETP ALL samples in the portal.
Here are the steps taken:
00. Pre-processing the provided SCE objects
01. Feature selection/dimensionality reduction
02. Cell type annotation with marker genes
03. Tumor cell identification
What software will you require?
Seurat,reticulate,SAM,ScType,CopyKatWhat will your first pull request contain?
./create-analysis-module.py cell-type-nonETP-ALL-03 --use-r --use-renv --use-conda --conda-file-only00. Pre-processing the provided SCE objectsWhat computational resources will you require?
I am currently using our own computing system for the analysis. I may need help in transferring files or large objects, if there is any issue encountered in the future.
If known, when do you expect to file the first pull request?
~ 09/20/2024