Merge pull request #789 from UTSouthwesternDSSR/main

celltype/tumor annotation for non-ETP T-ALL (SCPCP000003)

.github/workflows/docker_cell-type-nonETP-ALL-03.yml

@@ -13,22 +13,22 @@ concurrency:
   cancel-in-progress: true
 
 on:
-  # pull_request:
-  #   branches:
-  #     - main
-  #   paths:
-  #     - "analyses/cell-type-nonETP-ALL-03/Dockerfile"
-  #     - "analyses/cell-type-nonETP-ALL-03/.dockerignore"
-  #     - "analyses/cell-type-nonETP-ALL-03/renv.lock"
-  #     - "analyses/cell-type-nonETP-ALL-03/conda-lock.yml"
-  # push:
-  #   branches:
-  #     - main
-  #   paths:
-  #     - "analyses/cell-type-nonETP-ALL-03/Dockerfile"
-  #     - "analyses/cell-type-nonETP-ALL-03/.dockerignore"
-  #     - "analyses/cell-type-nonETP-ALL-03/renv.lock"
-  #     - "analyses/cell-type-nonETP-ALL-03/conda-lock.yml"
+  pull_request:
+    branches:
+      - main
+    paths:
+      - "analyses/cell-type-nonETP-ALL-03/Dockerfile"
+      - "analyses/cell-type-nonETP-ALL-03/.dockerignore"
+      - "analyses/cell-type-nonETP-ALL-03/renv.lock"
+      - "analyses/cell-type-nonETP-ALL-03/conda-lock.yml"
+  push:
+    branches:
+      - main
+    paths:
+      - "analyses/cell-type-nonETP-ALL-03/Dockerfile"
+      - "analyses/cell-type-nonETP-ALL-03/.dockerignore"
+      - "analyses/cell-type-nonETP-ALL-03/renv.lock"
+      - "analyses/cell-type-nonETP-ALL-03/conda-lock.yml"
   workflow_dispatch:
     inputs:
       push-ecr:

.github/workflows/run_cell-type-nonETP-ALL-03.yml

@@ -19,20 +19,23 @@ concurrency:
 
 on:
   workflow_dispatch:
-  # workflow_call:
-  # pull_request:
-  #   branches:
-  #     - main
-  #   paths:
-  #     - analyses/cell-type-nonETP-ALL-03/**
-  #     - "!analyses/cell-type-nonETP-ALL-03/Dockerfile"
-  #     - "!analyses/cell-type-nonETP-ALL-03/.dockerignore"
-  #     - .github/workflows/run_cell-type-nonETP-ALL-03.yml
+  workflow_call:
+  pull_request:
+    branches:
+      - main
+    paths:
+      - analyses/cell-type-nonETP-ALL-03/**
+      - "!analyses/cell-type-nonETP-ALL-03/Dockerfile"
+      - "!analyses/cell-type-nonETP-ALL-03/.dockerignore"
+      - .github/workflows/run_cell-type-nonETP-ALL-03.yml
 
 jobs:
   run-module:
     if: github.repository_owner == 'AlexsLemonade'
     runs-on: ubuntu-latest
+    defaults:
+      run:
+        shell: bash -el {0}
 
     steps:
       - name: Checkout repo
@@ -47,16 +50,43 @@ jobs:
       - name: Set up pandoc
         uses: r-lib/actions/setup-pandoc@v2
 
+      - name: Install system dependencies
+        run: |
+          sudo apt-get install -y libcurl4-openssl-dev \
+            libhdf5-dev \
+            libglpk40 \
+            libxml2-dev \
+            libfontconfig1-dev \
+            libharfbuzz-dev \
+            libfribidi-dev \
+            libtiff5-dev
+
       - name: Set up renv
         uses: r-lib/actions/setup-renv@v2
         with:
           working-directory: ${{ env.MODULE_PATH }}
 
+      - name: Set up conda
+      # Note that this creates and activates an environment named 'test' by default
+        uses: conda-incubator/setup-miniconda@v3
+        with:
+          miniforge-version: latest
+
+      - name: Install conda-lock and activate locked conda environment
+        run: |
+          conda install conda-lock
+          conda-lock install --name openscpca-cell-type-nonETP-ALL-03 ${MODULE_PATH}/conda-lock.yml
+
       # Update this step as needed to download the desired data
       - name: Download test data
-        run: ./download-data.py --test-data --format SCE
+        run: |
+          ./download-data.py --projects SCPCP000003 --test-data --format SCE
+          ./download-results.py --projects SCPCP000003 --test-data --modules doublet-detection
 
       - name: Run analysis module
         run: |
           cd ${MODULE_PATH}
           # run module script(s) here
+          Rscript scripts/00-01_processing_rds.R
+          Rscript scripts/02-03_annotation.R
+          Rscript scripts/multipanel_plot.R

analyses/cell-type-nonETP-ALL-03/.gitignore

@@ -2,10 +2,6 @@
 /results/*
 !/results/README.md
 
-#Plots should not be committed
-/plots/*
-!/plots/.gitkeep
-
 # Ignore the scratch directory (but keep it present)
 /scratch/*
 !/scratch/.gitkeep

analyses/cell-type-nonETP-ALL-03/Azimuth_BM_level1.csv

@@ -1,4 +1,4 @@
-tissueType,cellName,geneSymbolmore1,geneSymbolmore2,fullName,ontologyID
+tissueType,cellName,ensembl_id_positive_marker,ensembl_id_negative_marker,fullName,ontologyID
 Immune system,B,"ENSG00000163534,ENSG00000132704,ENSG00000012124,ENSG00000138639,ENSG00000153064,ENSG00000156738,ENSG00000116191,ENSG00000104894,ENSG00000133789,ENSG00000105369",,B cell,CL_0000945
 Immune system,CD4 T,"ENSG00000172005,ENSG00000138795,ENSG00000168685,ENSG00000081059,ENSG00000227507,ENSG00000104660,ENSG00000198851,ENSG00000167286,ENSG00000160654,ENSG00000139193",,CD4 T cell,CL_0000624
 Immune system,CD8 T,"ENSG00000172116,ENSG00000153563,ENSG00000184613,ENSG00000167286,ENSG00000198851,ENSG00000160307,ENSG00000100450,ENSG00000160654,ENSG00000227191,ENSG00000271503",,CD8 T cell,CL_0000625
@@ -13,4 +13,6 @@ Immune system,Late Eryth,"ENSG00000196188,ENSG00000112212,ENSG00000204010,ENSG00
 Immune system,Plasma,"ENSG00000115884,ENSG00000222037,ENSG00000211640,ENSG00000048462,ENSG00000211673,ENSG00000240505,ENSG00000211685,ENSG00000167476,ENSG00000143297,ENSG00000243466",,Plasma cell,CL_0000786
 Immune system,Platelet,"ENSG00000150681,ENSG00000187699,ENSG00000088726,ENSG00000169704,ENSG00000163737,ENSG00000163736,ENSG00000153071,ENSG00000113140,ENSG00000176783,ENSG00000124491",,Platelet,CL_0000233
 Immune system,Stromal,"ENSG00000115461,ENSG00000047457,ENSG00000091513,ENSG00000011465,ENSG00000139329,ENSG00000164692,ENSG00000147571,ENSG00000041982,ENSG00000152583,ENSG00000112175",,Stromal cell,CL_0000499
+Immune system,Blast,"ENSG00000002586,ENSG00000173762,ENSG00000124766,ENSG00000177606,ENSG00000117632,ENSG00000123416,ENSG00000167286",,Blast cell,CL_0000055
 Immune system,Cancer,"ENSG00000026508,ENSG00000119888,ENSG00000141736,ENSG00000086205,ENSG00000111057,ENSG00000007062",,Cancer cell,CL_0001064
+Immune system,Pre Eryth,"ENSG00000081237,ENSG00000170180,ENSG00000175792,ENSG00000072274,ENSG00000110195,ENSG00000135218,ENSG00000115232,ENSG00000244734,ENSG00000223609,ENSG00000133742",,Erythroid-like and erythroid precursor cell,CL_0000038

analyses/cell-type-nonETP-ALL-03/Dockerfile

@@ -1,10 +1,65 @@
-# A template docker file for creating a new analysis
-FROM ubuntu:22.04
+
+FROM bioconductor/r-ver:3.19
 
 # Labels following the Open Containers Initiative (OCI) recommendations
 # For more information, see https://specs.opencontainers.org/image-spec/annotations/?v=v1.0.1
+LABEL org.opencontainers.image.title="openscpca/cell-type-nonETP-ALL-03"
+LABEL org.opencontainers.image.description="Docker image for the OpenScPCA analysis module 'cell-type-nonETP-ALL-03'"
 LABEL org.opencontainers.image.authors="OpenScPCA [email protected]"
 LABEL org.opencontainers.image.source="https://github.com/AlexsLemonade/OpenScPCA-analysis/tree/main/templates/analysis-module"
 
 # Set an environment variable to allow checking if we are in an OpenScPCA container
 ENV OPENSCPCA_DOCKER=TRUE
+
+# set a name for the conda environment
+ARG ENV_NAME=openscpca-cell-type-nonETP-ALL-03
+
+# set environment variables to install conda
+ENV PATH="/opt/conda/bin:${PATH}"
+
+# Install conda via miniforge
+# adapted from https://github.com/conda-forge/miniforge-images/blob/master/ubuntu/Dockerfile
+RUN curl -L "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh" -o /tmp/miniforge.sh \
+  && bash /tmp/miniforge.sh -b -p /opt/conda \
+  && rm -f /tmp/miniforge.sh \
+  && conda clean --tarballs --index-cache --packages --yes \
+  && find /opt/conda -follow -type f -name '*.a' -delete \
+  && find /opt/conda -follow -type f -name '*.pyc' -delete \
+  && conda clean --force-pkgs-dirs --all --yes
+
+# Activate conda environments in bash
+RUN ln -s /opt/conda/etc/profile.d/conda.sh /etc/profile.d/conda.sh \
+  && echo ". /opt/conda/etc/profile.d/conda.sh" >> /etc/skel/.bashrc \
+  && echo ". /opt/conda/etc/profile.d/conda.sh" >> ~/.bashrc
+
+# Install conda-lock
+RUN conda install --channel=conda-forge --name=base conda-lock \
+  && conda clean --all --yes
+
+# Install renv
+RUN Rscript -e "install.packages('renv')"
+
+# Disable the renv cache to install packages directly into the R library
+ENV RENV_CONFIG_CACHE_ENABLED=FALSE
+
+# Copy conda lock file to image
+COPY conda-lock.yml conda-lock.yml
+
+# restore from conda-lock.yml file and clean up to reduce image size
+RUN conda-lock install -n ${ENV_NAME} conda-lock.yml \
+  && conda clean --all --yes
+
+# Copy the renv.lock file from the host environment to the image
+COPY renv.lock renv.lock
+
+# restore from renv.lock file and clean up to reduce image size
+RUN Rscript -e 'renv::restore()' \
+  && rm -rf ~/.cache/R/renv \
+  && rm -rf /tmp/downloaded_packages \
+  && rm -rf /tmp/Rtmp*
+
+# Activate conda environment on bash launch
+RUN echo "conda activate ${ENV_NAME}" >> ~/.bashrc
+
+# Set CMD to bash to activate the environment when launching
+CMD ["/bin/bash"]

analyses/cell-type-nonETP-ALL-03/README.md

@@ -1,12 +1,12 @@
 # Non-ETP T-ALL Annotation (SCPCP000003)
 
-This analysis module will include code to annotate cell types in non-ETP T-ALL from SCPCP000003 (n=11) present on the ScPCA portal.
+This analysis module will include code to annotate cell types and tumor/normal status in non-ETP T-ALL from SCPCP000003 (n=11) present on the ScPCA portal.
 
 ## Description
 
 We first aim to annotate the cell types in non-ETP T-ALL, and use the annotated B cells in the sample as the "normal" cells to identify tumor cells, since T-ALL is caused by the clonal proliferation of immature T-cell [<https://www.nature.com/articles/s41375-018-0127-8>].
 
--   We use the cell type marker (`Azimuth_BM_level1.csv`) from [Azimuth Human Bone Marrow reference](https://azimuth.hubmapconsortium.org/references/#Human%20-%20Bone%20Marrow). In total, there are 14 cell types: B, CD4T, CD8T, Other T, DC, Monocytes, Macrophages, NK, Early Erythrocytes, Late Erythrocytes, Plasma, Platelet, Stromal, and Hematopoietic Stem and Progenitor Cells (HSPC). Based on the exploratory analysis, we believe that most of the cells in these samples do not express adequate markers to be distinguished at finer cell type level (eg. naive vs memory, CD14 vs CD16 etc.), and majority of the cells should belong to T-cells. In addition, we include the marker genes for cancer cell in immune system from [ScType](https://sctype.app/database.php) database.
+-   We use the cell type marker (`Azimuth_BM_level1.csv`) from [Azimuth Human Bone Marrow reference](https://azimuth.hubmapconsortium.org/references/#Human%20-%20Bone%20Marrow). In total, there are 14 cell types: B, CD4T, CD8T, Other T, DC, Monocytes, Macrophages, NK, Early Erythrocytes, Late Erythrocytes, Plasma, Platelet, Stromal, and Hematopoietic Stem and Progenitor Cells (HSPC). Based on the exploratory analysis, we believe that most of the cells in these samples do not express adequate markers to be distinguished at finer cell type level (eg. naive vs memory, CD14 vs CD16 etc.), and majority of the cells should belong to T-cells. In addition, we include the marker genes for blast cell [[Bhasin et al. (2023)](https://www.nature.com/articles/s41598-023-39152-z)] as well as erythroid precursor and cancer cell in immune system [[ScType](https://sctype.app/database.php) database].
 
 -   Since ScType annotates cell types at cluster level using marker genes provided by user or from the built-in database, we employ [self-assembling manifold](https://github.com/atarashansky/self-assembling-manifold/tree/master) (SAM) algorithm, a soft feature selection strategy for better separation of homogeneous cell types.
 
@@ -42,7 +42,7 @@ The `scripts/00-01_processing_rds.R` requires the processed SingleCellExperiment
 ../../download-results.py --projects SCPCP000003 --modules doublet-detection
 ```
 
-As for the annotation, `scripts/02-03_annotation.R` requires cell type marker gene file, `Azimuth_BM_level1.csv`, as an input for ScType. This excel file contains a list of positive marker genes in Ensembl ID under `geneSymbolmore1` for each cell type, and *TMEM56* is not detected in our dataset, thus it is being removed as part of the markers for Late Eryth. As of now, there is no negative marker genes provided under `geneSymbolmore2`.
+As for the annotation, `scripts/02-03_annotation.R` requires cell type marker gene file, `Azimuth_BM_level1.csv`, as an input for ScType. This excel file contains a list of positive marker genes in Ensembl ID under `ensembl_id_positive_marker` for each cell type; *TMEM56* and *CD235a* are not detected in our dataset, thus they are being removed as part of the markers for Late Eryth and Pre Eryth respectively. As of now, there is no negative marker genes provided under `ensembl_id_negative_marker`.
 
 ## Output files
 
@@ -52,6 +52,14 @@ Running `scripts/00-01_processing_rds.R` will generate two types of output:
 
 -   umap plots showing leiden clustering in `plots/`
 
+Running `scripts/02-03_annotation.R` will generate several outputs:
+
+-   updated `rds` objects in `scratch/`
+
+-   umap plots showing cell type and CopyKat prediction (if there is any) and dotplots showing the features added with `AddModuleScore()` in `plots/`
+
+-   ScType results of top 10 possible cell types in a cluster (`_sctype_top10_celltypes_perCluster.txt`) and metadata file tabulating leiden cluster, cell type, low confidence cell type, and CopyKat prediction for each cell (`_metadata.txt`) in `results/`
+
 ## Software requirements
 
 To run the analysis, execute the Rscript in R or Rstudio (version 4.4.0). The main libraries used are: