Merge branch 'master' of github.com:Acribbs/scflow

README.md

@@ -14,26 +14,20 @@ This repository contains a collection of pipelines that aid the analysis of sing
 
 ## Installation
 
-### pip install
-
-You can install scflow using pip, this will only install the package without any dependancies, which will have to be installed seperately.::
-
-	pip install scflow
-
 ### Conda installation - **in progress**
 
-The preferred method for installation is through conda. Currently this installation is still in working progress. Preferably the
+The preferred method for installation is through conda/mamba.  Preferably the
 installation should be in a seperate environment::
 
-    conda create -n scflow -c cgat scflow
+    mamba env create -f conda/environments/scflow.yml
     conda activate scflow
-    scflow --help
-
-### Manual installation
+    python setup.py develop
 
-The repository can also be installed manually, but dependencies will need to be installed seperately::
-
-    python setup.py install
+    # Install a specific version of kb-tools from Adam's cloned repo
+    git clone git@github.com:Acribbs/kb_python.git
+    cd kb_python
+    python setup.py develop
+
     scflow --help
 
 ## Usage
@@ -86,6 +80,10 @@ code can be found at [read the docs](http://single-cell.readthedocs.io/)
 
 - [ ] [Overview of the seurat doublet-4 pipeline](docs/pipelines/seurat_doublet-4.md)
 
+## seurat integration-5
+
+- [ ] [Overview of the seurat doublet-4 pipeline](docs/pipelines/seurat_doublet-4.md)
+
 # Project Info
 
 - [ ] [Contributors](docs/project_info/Contributing.rst)

conda/environments/scflow.yml

@@ -26,17 +26,24 @@ dependencies:
  - setuptools
  - statsmodels
  - r-base
- - bioconductor-biomart
- - bioconductor-singlecellexperiment
  - r-optparse
  - r-seurat
  - r-tidyverse
  - r-ggthemes
- - bioconductor-scdblfinder
  - scanpy
  - scipy
  - papermill
  - leidenalg
  - bioconductor-busparse 
  - bustools
  - bioconductor-tximport
+ - bioconductor-dropletutils
+ - bioconductor-celda
+ - bioconductor-scdblfinder
+ - bioconductor-singlecellexperiment
+ - bioconductor-biomart
+ - bioconductor-scuttle
+ - bioconductor-glmgampoi
+ - r-harmony
+ - r-seuratdisk
+ - r-clustree

docs/pipelines/seurat_integration-5.md

@@ -0,0 +1,51 @@
+## seurat integration-5
+
+The pipeline follows pipeline quantnuclei, qc-1, filter-2 and optionally cluster-3.
+
+It takes filtered Seurat objects and performs integrations by two methods, Seurat and Harmony.
+Each dataset is first processed individually by SCTransform normalization and then perform integration and clustering by Seurat.
+The integrated data could be furthered accepted by Harmony with integration/clustering.
+The integrations performed by both the methods could be visualized as tSNE and UMAP dimensional reduction.
+
+**Overview**  
+Runs the R script "seurat_integrate.R"  
+Runs the R script "harmony_integrate.R"  
+Runs Rmd file "Integration.Rmd" to visualise the results.
+
+**Commands:**  
+
+Configure the pipeline.yml file
+
+> scflow seurat doublet-4 config  
+
+Run the pipeline
+> nohup scflow seurat doublet-4 make full -v5
+
+### seurat_integrate.R
+
+**Inputs:**  
+Filtered clustered seurat object.  
+(Or would work using filtered seurat object from output of filter-2.
+
+**Steps:**  
+- Read in the .yml file
+- Extract parameters from the .yml file
+- Read in the samples
+- Perform integration steps (creates a new seurat object which contains the integrated data):
+    - Normalise the data by SCTransform
+    - Select variable features that are common across the Seurat Objects
+    - Use PrepSCTIntegration to get residuals for all features
+    - Identify anchors
+    - Integrate the data  
+- Perform PCA
+- Find Neighbours
+- Find Clusters
+- Perform UMAP
+- Plot and save UMAPs
+- Perform tSNE
+- Plot and save tSNEs
+- Save integrated object as RDS file
+
+**Outputs**
+UMAP and tSNE Plots  
+Integrated RDS object  

docs/pipelines/seurat_qc-1.md

@@ -17,26 +17,33 @@ The pipeline pipeline_qc-1.py runs an R markdown file called QC.Rmd to assess th
 **Inputs:**
 
 Count matrix generated from quantnuclei pipeline  
+Patient metadata as tab-delimited file    
+
 
 **Steps:**
 1. Read in the count matrix
-2. Create Seurat object, Normalize data, scale data and find variable features
-3. Create some ggplot themes
-4. Add columns to metadata: nUMI, nGene, log10GenesPerUMI
-5. Identify mitochondrial genes and add to metadata
-6. Map ensembl symbols to hgnc symbols using biomart
-7. Filter out genes and cells with low number of counts
-8. Create SingleCellExperiment object and save RDS files
+2. Create Seurat object
+3. Generate additional QC metrics: percent mitochondrial genes, log10GenesPerUMI
+4. Save the mapping reference for ensemble gene names to gene IDs
+5. Add gene symbols to the meta.features of the RNA assay
+6. Add patient metadata
+7. Convert to SingleCellExperiment 
+	- Identify and remove empty droplets (NB empty droplets must be removed in order to perform doublet detection).  
+	- Identify ambient RNA
+	- Identify doublets
+8. Save Seurat Objects and SingleCellExperiment objects as RDS files (both complete and with emptry droplets filtered out)
 9. Plot and save QC metrics
 	- Cell counts per sample
 	- UMI counts per cell
 	- Genes detected per cell
-	- UMIs vs genes detected
 	- Mitochondrial counts ratio
+	- UMIs vs genes detected
 	- Novelty
 
 **Outputs:**
 
-QC.Rmd knitted to html
-SingleCellExperiment and Seurat Object RDS objects saved in RDS_objects.dir
-QC plots saved as .eps files in QC_Figures.dir
+QC.Rmd knitted to html  
+
+SingleCellExperiment and Seurat Object RDS objects saved in RDS_objects.dir/unfiltered    
+QC plots saved as .png files in QC_Figures.dir  
+mapping.txt saved in Files.dir  

scpipelines/R/plot_barnyard.R

@@ -8,7 +8,7 @@ library(optparse)
 
 option_list = list(
   make_option(c("-i", "--input"), type="character", default="", 
-              help="input file GTF file [default= %default]", metavar="character"),
+              help="input file path for mtx gene matrix location [default= %default]", metavar="character"),
   make_option(c("-o", "--out"), type="character", default="", 
               help="output file name [default= %default]", metavar="character")
 	      );