bioinfo core repo with aws setup instructions

ucdavis-bioinformatics · Oct 21, 2019 · b5ac87e · b5ac87e
1 parent c9e2957
commit b5ac87e
Show file tree

Hide file tree

Showing 18 changed files with 319 additions and 1 deletion.
diff --git a/.DS_Store b/.DS_Store
diff --git a/README.md b/README.md
@@ -1 +1 @@
-
+# scRNA_shiny_app
diff --git a/app.R b/app.R
@@ -0,0 +1,251 @@
+library(shiny)
+library(Seurat)
+library(ggplot2)
+library(tidyr)
+library(dplyr)
+
+load('experiment_merged.RData')
+genes = experiment.merged@assays$RNA
+meta_nums <- colnames(dplyr::select_if(experiment.merged@meta.data, is.numeric))
+meta_cats <- colnames(dplyr::select_if(experiment.merged@meta.data, is.factor))
+pcs <- list('PC_1','PC_2','PC_3','PC_4','PC_5','PC_6','PC_7','PC_8','PC_9')
+agg_cats <- colnames(dplyr::select_if(experiment.merged@meta.data, is.factor))
+
+
+server = function(input, output, session){
+  outVar = reactive({
+    if (input$dataset == 'Genes'){mydata=row.names(genes)}
+    else if (input$dataset == 'Numeric Metadata') {mydata=meta_nums}
+    else if (input$dataset == 'PCs') {mydata=pcs}
+    mydata
+  })
+
+  outVar = reactive({
+    if (input$dataset_single == 'Genes'){mydata=row.names(genes)}
+    else if (input$dataset_single == 'Numeric Metadata') {mydata=meta_nums}
+    else if (input$dataset_single == 'PCs') {mydata=pcs}
+    mydata
+  })
+
+  observe({
+    updateSelectInput(session, "numeric",
+                      choices = outVar()
+  )})
+
+  observe({
+    updateSelectInput(session, "numeric2",
+                      choices = outVar()
+  )})
+
+  observe({
+    updateSelectInput(session, "numeric_b",
+                      choices = outVar()
+  )})
+
+  observe({
+    updateSelectInput(session, "numeric_single",
+                      choices = outVar()
+    )})
+
+  # formulaText <- reactive({
+  #   paste("Marker Gene ~", input$numeric)
+  # })
+
+  # Marker Plot Double
+  output$MarkerGenePlot <- renderPlot({
+    FeaturePlot(
+      experiment.merged,
+      c(input$numeric, input$numeric2), blend=TRUE
+    )
+  })
+
+  # Marker Plot Single
+  output$MarkerGenePlotSingle <- renderPlot({
+    FeaturePlot(
+      experiment.merged,
+      c(input$numeric_single)
+    )
+  })
+
+  # Double Feature Categorical Feature Plot
+  output$CategoricalPlot <- renderPlot({
+    DimPlot(object = experiment.merged, group.by=input$categorical, pt.size=0.5, do.label = TRUE, reduction = "tsne", label = T)
+  })
+
+  # Single Feature Categorical Feature Plot
+  output$CategoricalPlotSingle <- renderPlot({
+    DimPlot(object = experiment.merged, group.by=input$categorical_single, pt.size=0.5, do.label = TRUE, reduction = "tsne", label = T)
+  })
+
+  # Double Feature Violin Plot
+  output$ViolinPlot <- renderPlot({
+    Idents(experiment.merged) <- input$categorical
+    VlnPlot(object =  experiment.merged, features = c(input$numeric, input$numeric2), pt.size = 0.05)
+  })
+
+  # Single Feature Violin Plot
+  output$ViolinPlotSingle <- renderPlot({
+    Idents(experiment.merged) <- input$categorical
+    VlnPlot(object =  experiment.merged, features = c(input$numeric_single), pt.size = 0.05)
+  })
+
+  # Marker Set Plot
+  output$MarkerSet <- renderPlot({
+    Idents(experiment.merged) <- input$categorical_b
+    markers = input$numeric_b
+    expr.cutoff = 3
+    widedat <- FetchData(experiment.merged, markers)
+    widedat$Cluster <- Idents(experiment.merged)
+    longdat <- gather(widedat, key = "Gene", value = "Expression", -Cluster)
+    longdat$Is.Expressed <- ifelse(longdat$Expression > expr.cutoff, 1, 0)
+    longdat$Cluster <- factor(longdat$Cluster)
+    longdat$Gene <- factor(longdat$Gene)
+
+    # Need to summarize into average expression, pct expressed (which is also an average)
+    plotdat <- group_by(longdat, Gene, Cluster) %>% summarize(`Percentage of Expressed Cells` = mean(Is.Expressed), `Mean Expression` = mean(Expression))
+    ggplot(plotdat, aes(x = Gene, y = Cluster)) +
+      geom_point(aes(size = `Percentage of Expressed Cells`, col = `Mean Expression`)) +
+      labs(size = "Percentage\nof Expressed\nCells", col = "Mean\nExpression", x = NULL) +
+      scale_color_gradient(low = "grey", high = "slateblue4") + theme_grey(base_size = 15) +
+      theme(axis.text.x = element_text(angle = 90, hjust = 1))
+  # }, height = 1000, width = 900 )
+  }, height = 1000)
+
+}
+
+# ui <- fluidPage(
+#   
+#   # App title ----
+#   titlePanel("scRNA Seurat Analysis"),
+#   
+#   # Sidebar layout with input and output definitions ----
+#   sidebarLayout(
+#     
+#     # Sidebar panel for inputs ----
+#     sidebarPanel(
+#         conditionalPanel(condition = "input.tabselected == -999",
+#                 selectInput("dataset", "Numeric Analysis Type:",
+#                             c('Genes', 'Numeric Metadata','PCs')),
+#                 selectInput("categorical", "Identity:",
+#                             c(meta_cats)),
+#                 selectInput("numeric", "Primary Numeric:", ""),
+# 
+#                 selectInput('numeric2', 'Secondary Numeric', "")
+#         ),
+# 
+#         conditionalPanel(condition = "input.tabselected == 2",
+#                          selectInput("categorical_b", "Identity:",
+#                                      c(agg_cats)),
+#                          selectInput("numeric_b", "Primary Numeric:", "", multiple=TRUE)
+#          ),
+#         conditionalPanel(condition = "input.tabselected == 3")
+#     ),
+# 
+#     # Main panel for displaying outputs ----
+#     mainPanel(
+#       
+#       # Output: Tabset w/ plot, summary, and table ----
+#       navbarPage("My application",
+#                   tabPanel("Marker Genes (TSNE)", value=-999,
+#                                #h3(textOutput("caption")),
+#                                plotOutput("MarkerGenePlot"),
+#                                plotOutput("ViolinPlot"),
+#                                #h3(textOutput("caption2")),
+#                                plotOutput("CategoricalPlot")
+#                                #h3(textOutput("caption3")),
+#                            ),
+#                   
+#                   tabPanel("Marker Set (Grid)", value=2,
+#                                plotOutput("MarkerSet")
+#                           ),
+# 
+#                   tabPanel("Documentation", value=3,
+#                            includeMarkdown("docs/testing.md"),
+#                            includeMarkdown("docs/todo.md")
+#                          ),
+#                   
+#                   id = "tabselected"
+#       )
+#     )
+#   )
+# )
+
+
+
+ui <- fluidPage(
+
+  titlePanel("scRNA Seurat Analysis"), 
+  sidebarLayout(
+    sidebarPanel(width = 12,
+                 tabsetPanel(
+                   tabPanel("Documentation", value=-999,
+                            mainPanel(width = 12,
+                              br(),
+                              includeMarkdown("docs/testing.md"),
+                              includeMarkdown("docs/todo.md")
+                            )
+                   ),
+
+                   tabPanel("Double Marker", value=2,
+                                br(),
+                                div(style="display: inline-block;vertical-align:top; width: 24%;",
+                                    selectInput("dataset", "Numeric Analysis Type:",
+                                            c('Genes', 'Numeric Metadata','PCs'))),
+                                div(style="display: inline-block;vertical-align:top; width: 24%;",
+                                    selectInput("categorical", "Identity:",
+                                            c(meta_cats))),
+                                div(style="display: inline-block;vertical-align:top; width: 24%;",
+                                    selectInput("numeric", "Primary Numeric:", "")),
+
+                                div(style="display: inline-block;vertical-align:top; width: 24%;",
+                                    selectInput('numeric2', 'Secondary Numeric', "")),
+
+                              mainPanel(width = 12,
+                                 br(),
+                                 br(),
+                                 #h3(textOutput("caption")),
+                                 plotOutput("MarkerGenePlot"),
+                                 plotOutput("ViolinPlot"),
+                                 plotOutput("CategoricalPlot")
+                              )
+                          ),
+                   tabPanel("Single Marker", value=3,
+                            br(),
+                            div(style="display: inline-block;vertical-align:top; width: 24%;",
+                            selectInput("dataset_single", "Numeric Analysis Type:",
+                                        c('Genes', 'Numeric Metadata','PCs'))),
+                            div(style="display: inline-block;vertical-align:top; width: 24%;",
+                                selectInput("categorical_single", "Identity:",
+                                            c(meta_cats))),
+                            div(style="display: inline-block;vertical-align:top; width: 24%;",
+                                selectInput("numeric_single", "Primary Numeric:", "")),
+
+                            mainPanel(width = 12,
+                                      br(),
+                                      br(),
+                                      #h3(textOutput("caption")),
+                                      plotOutput("MarkerGenePlotSingle"),
+                                      plotOutput("ViolinPlotSingle"),
+                                      plotOutput("CategoricalPlotSingle")
+                            )
+                   ),
+                  tabPanel("Marker Set (Grid)", value=4,
+                            br(),
+                            selectInput("categorical_b", "Identity:",
+                                        c(agg_cats)),
+                            selectInput("numeric_b", "Primary Numeric:", "", multiple=TRUE),
+                             mainPanel(width = 12,
+                                br(),
+                                br(),
+                                plotOutput("MarkerSet")
+                             )
+                          ),
+                  id = "tabselected"
+                 )
+    ),
+    mainPanel(width = 12)
+  )
+)
+
+
+shinyApp(ui, server)
diff --git a/aws_setup/README.md b/aws_setup/README.md
diff --git a/aws_setup/start_app.md b/aws_setup/start_app.md
@@ -0,0 +1,38 @@
+# Starting a Shiny Server Instance from the Preconfigured Image
+
+### Keith Mitchell (kgmitchell@ucdavis.edu)
+***
+#### A few general notes:
+- be sure Adam has granted you an IAM account on the UC Davis Bioinformatics AWS account. 
+- Do not nuke the Zulip instances!
+***
+#### How to prepare the instance for a client:
+1. Go to the AWS AMI's (Amazon Machine Images). Select the Rshiny Server Image. Go to Actions>Launch
+![](./start_app1.png)
+2. Next Choose and Instance Type. This will vary in general based on the size of the data you are attempting to analyze. 
+In general it is advised to start with a smaller instance and then move up from there if necessary.
+![](./start_app2.png)
+3. Next Configure the Instance Details. The main step here is to set the "Auto-assign Public IP" to enable.
+![](./start_app3.png)
+4. Keep clicking the bottom right button, "Next: Configure Security Group" until you get to "Step 6: Configure Security Group"
+5. At step 6, select the Rshiny App security group.
+![](./start_app4.png)
+6. Finally launch the instance. (Bottom right hand quarter)
+![](./start_app5.png)
+7. Select an existing key pair or create a new key pair is the final step for starting the instance. 
+Choose "Create a new key pair" and give the key a useful name, here we will use "tutorial".
+![](./start_app6.png)
+8. Move the downloaded key to `~/.ssh/` via `mv ~/Downloads/tutorial.pem ~/.ssh/`
+9. `chmod 700 ~/.ssh/tutorial.pem `
+10. Get the connection to the instance seen highlighted below "ec2-…….amazonaws.com"
+![](./start_app7.png)
+11. Run the following command to upload the "experiment_merged.Rdata" object to the app:
+    - `scp -i ~/.ssh/tutorial.pem ec2-54-219-166-77.us-west-1.compute.amazonaws.com:/srv/shiny-server/scRNA_shiny_app/  ~/Desktop/experiment_merged.Rdata`
+13. Replace the green section with your Public IV4 DNS address from the instance that was created and used in the previous step.
+    - http://**ec2-54-219-166-77.us-west-1.compute.amazonaws.com**:3838/scRNA_shiny_app/
+14. If there are problems with this link do the following:
+    - SSH into the server.
+        - `ssh -i ~/.ssh/tutorial.key ec2-54-219-166-77.us-west-1.compute.amazonaws.com`
+    - Restart the shiny service.
+        - `sudo systemctl restart shiny-server.service`
+
diff --git a/aws_setup/start_app1.png b/aws_setup/start_app1.png
diff --git a/aws_setup/start_app2.png b/aws_setup/start_app2.png
diff --git a/aws_setup/start_app3.png b/aws_setup/start_app3.png
diff --git a/aws_setup/start_app4.png b/aws_setup/start_app4.png
diff --git a/aws_setup/start_app5.png b/aws_setup/start_app5.png
diff --git a/aws_setup/start_app6.png b/aws_setup/start_app6.png
diff --git a/aws_setup/start_app7.png b/aws_setup/start_app7.png
diff --git a/docs/testing.md b/docs/testing.md
diff --git a/docs/todo.md b/docs/todo.md
@@ -0,0 +1,29 @@
+# Todo:
+- readme documentation tab with the primers/library used
+- change the colors from (green and red) hack the Feature plot function seurat:::function
+- dont hard code the PCs only grab the first 10
+
+# Done or not needed
+## newer
+- stretch out the 2nd plot
+- make the docs the first tab
+- layout all the way across the top
+- make NA an option for the second gene (binary button)(kinda done? just did a variation of this)
+
+
+
+## older
+- ensembl gene id's symbol if available 
+- add an option for the PC graphs 
+- marker set tab with the figure Blythe/Matt had seen
+- readme tab (MD)
+- make the side by side (secondary gene) optional
+    - going along with this Matt wanted to change the layout.. I did put the violin plot in the middle
+    - not sure i can group them around the side bar layout as wanted. 
+
+
+# Not sure how to do it
+- make the default NA
+
+# Questions
+- is it ok to use experiment merged for everything (has more groupings) OK
diff --git a/experiment_merged.RData b/experiment_merged.RData
diff --git a/original_seurat_object.RData b/original_seurat_object.RData
diff --git a/outs/.DS_Store b/outs/.DS_Store
diff --git a/pca_sample_corrected.RData b/pca_sample_corrected.RData