refactor: Make changes from feedback from HUPO (#26)

* layout(visualizeNetworks): Set up cose layout for network viz

* refactor(getSubnetworkFromIndra): Add input validation

* fix visualizeNetwork tests

* Add validation tests for hgncID column being missing

NAMESPACE

@@ -5,6 +5,7 @@ export(getSubnetworkFromIndra)
 export(visualizeNetworks)
 importFrom(RCy3,createNetworkFromDataFrames)
 importFrom(RCy3,createVisualStyle)
+importFrom(RCy3,layoutNetwork)
 importFrom(RCy3,mapVisualProperty)
 importFrom(RCy3,setVisualStyle)
 importFrom(httr,POST)

R/annotateProteinInfoFromIndra.R

@@ -2,7 +2,7 @@
 #'
 #' This function annotates a data frame with protein information from Indra.
 #'
-#' @param input output of \code{\link[MSstats]{groupComparison}} function's 
+#' @param df output of \code{\link[MSstats]{groupComparison}} function's 
 #'      comparisonResult table, which contains a list of proteins and their 
 #'      corresponding p-values, logFCs, along with additional HGNC ID and HGNC 
 #'      name columns

R/getSubnetworkFromIndra.R

@@ -25,6 +25,7 @@
 #'
 getSubnetworkFromIndra <- function(input, pvalueCutoff = NULL) {
     input <- .filterGetSubnetworkFromIndraInput(input, pvalueCutoff)
+    .validateGetSubnetworkFromIndraInput(input)
     res <- .callIndraCogexApi(input$HgncId)
     nodes <- .constructNodesDataFrame(input)
     edges <- .constructEdgesDataFrame(res, input)

R/utils_getSubnetworkFromIndra.R

@@ -1,3 +1,16 @@
+#' Validate input for MSstatsBioNet getSubnetworkFromIndra
+#' @param input dataframe from MSstats groupComparison output
+#' @keywords internal
+#' @noRd
+.validateGetSubnetworkFromIndraInput <- function(input) {
+    if (!"HgncId" %in% colnames(input)) {
+        stop("Invalid Input Error: Input must contain a column named 'HgncId'.")
+    }
+    if (nrow(input) >= 400) {
+        stop("Invalid Input Error: INDRA query must contain less than 400 proteins.  Consider adding a p-value cutoff")
+    }
+}
+
 #' Call INDRA Cogex API and return response
 #' @param hgncIds list of hgnc ids
 #' @return list of INDRA statements

R/visualizeNetworks.R

@@ -14,7 +14,7 @@
 #' @param logfcCutoff log fold change cutoff for coloring significant
 #' proteins. Default is 0.5
 #' @importFrom RCy3 createNetworkFromDataFrames mapVisualProperty
-#' createVisualStyle setVisualStyle
+#' createVisualStyle setVisualStyle layoutNetwork
 #'
 #' @export
 #'
@@ -30,7 +30,8 @@
 #'
 #'
 visualizeNetworks <- function(nodes, edges,
-                              pvalueCutoff = 0.05, logfcCutoff = 0.5) {
+                              pvalueCutoff = 0.05, 
+                              logfcCutoff = 0.5) {
     # Add additional columns for visualization
     nodes$logFC_color <- nodes$logFC
     nodes$logFC_color[nodes$pvalue > pvalueCutoff |
@@ -55,7 +56,7 @@ visualizeNetworks <- function(nodes, edges,
             mapVisualProperty(
                 "Node Fill Color", "logFC_color", "c",
                 c(-logfcCutoff, 0.0, logfcCutoff),
-                c("#5588DD", "#5588DD", "#D3D3D3", "#DD8855", "#DD8855")
+                c("#ADD8E6", "#ADD8E6", "#D3D3D3", "#FFA590", "#FFA590")
             )
         )
         createVisualStyle(
@@ -64,6 +65,7 @@ visualizeNetworks <- function(nodes, edges,
             VISUAL_STYLE_MAPPINGS
         )
         setVisualStyle(VISUAL_STYLE_NAME)
+        layoutNetwork("cose")
     } else {
         warning("Visualization is not available in non-interactive mode.")
     }

tests/testthat/test-getSubnetworkFromIndra.R

@@ -23,3 +23,28 @@ test_that("getSubnetworkFromIndra with pvalue filter works correctly", {
     expect_equal(nrow(subnetwork$nodes), 6)
     expect_equal(nrow(subnetwork$edges), 1)
 })
+
+test_that("Exception is thrown for 400+ proteins in dataframe", {
+    input_400 <- data.frame(
+        Protein = paste0("Protein", 1:400),
+        HgncId = paste0("HGNCID", 1:400),
+        issue = NA,
+        adj.pvalue = 0.05
+    )
+    expect_error(
+        getSubnetworkFromIndra(input_400),
+        "Invalid Input Error: INDRA query must contain less than 400 proteins.  Consider adding a p-value cutoff"
+    )
+})
+
+test_that("Exception is thrown for missing columns in input", {
+    input_missing_hgnc_id <- data.frame(
+        Protein = paste0("Protein", 1:10),
+        issue = NA,
+        adj.pvalue = 0.05
+    )
+    expect_error(
+        getSubnetworkFromIndra(input_missing_hgnc_id),
+        "Invalid Input Error: Input must contain a column named 'HgncId'."
+    )
+})

tests/testthat/test-visualizeNetworks.R

@@ -28,12 +28,18 @@ test_that("visualizeNetworks works correctly", {
         visualizeNetworks, "setVisualStyle",
         mock_setVisualStyle
     )
+    mock_layoutNetwork <- mock()
+    stub(
+        visualizeNetworks, "layoutNetwork",
+        mock_layoutNetwork
+    )
 
     expect_silent(visualizeNetworks(input$nodes, input$edges))
     expect_called(mock_createNetworkFromDataFrames, 1)
     expect_called(mock_mapVisualProperty, 2)
     expect_called(mock_createVisualStyle, 1)
     expect_called(mock_setVisualStyle, 1)
+    expect_called(mock_layoutNetwork, 1)
 })
 
 
@@ -67,6 +73,11 @@ test_that("visualizeNetworks with p-value and logFC constraints works", {
         visualizeNetworks, "setVisualStyle",
         mock_setVisualStyle
     )
+    mock_layoutNetwork <- mock()
+    stub(
+        visualizeNetworks, "layoutNetwork",
+        mock_layoutNetwork
+    )
 
     expect_silent(visualizeNetworks(input$nodes, input$edges,
         pvalueCutoff = 0.01, logfcCutoff = 2.5