Filter by PTM site match

R/getSubnetworkFromIndra.R

@@ -32,7 +32,9 @@
 #' as "namespace:identifier", e.g. "HGNC:1234" or "CHEBI:4911".
 #' @param filter_by_curation logical, whether to filter out statements that
 #' have been curated as incorrect in INDRA.  Default is FALSE.
-#' @param api_key string of INDRA API key for accessing curated statements.
+#' @param filter_by_ptm_site logical, whether to filter edges based on whether the 
+#' site information from INDRA matches with the PTM site in the input.  Default is FALSE.  
+#' Only applicable for differential PTM abundance results.
 #'
 #' @return list of 2 data.frames, nodes and edges
 #'
@@ -57,21 +59,23 @@ getSubnetworkFromIndra <- function(input,
                                    sources_filter = NULL,
                                    logfc_cutoff = NULL,
                                    force_include_other = NULL, 
-                                   filter_by_curation = FALSE, 
-                                   api_key = "") {
+                                   filter_by_curation = FALSE,
+                                   filter_by_ptm_site = FALSE) {
     input <- .filterGetSubnetworkFromIndraInput(input, pvalueCutoff, logfc_cutoff, force_include_other)
     .validateGetSubnetworkFromIndraInput(input, protein_level_data, sources_filter, force_include_other)
     res <- .callIndraCogexApi(input$HgncId, force_include_other)
-    res <- .filterIndraResponse(res, statement_types, evidence_count_cutoff, sources_filter, filter_by_curation, api_key)
+    res <- .filterIndraResponse(res, statement_types, evidence_count_cutoff, sources_filter)
     edges <- .constructEdgesDataFrame(res, input, protein_level_data)
     edges <- .filterEdgesDataFrame(edges, paper_count_cutoff, correlation_cutoff)
     nodes <- .constructNodesDataFrame(input, edges)
+    subnetwork = .filterByPtmSite(nodes, edges, filter_by_ptm_site)
+    subnetwork = .filterByCuration(subnetwork$nodes, subnetwork$edges, evidence_count_cutoff, filter_by_curation)
     warning(
         "NOTICE: This function includes third-party software components
         that are licensed under the BSD 2-Clause License. Please ensure to
         include the third-party licensing agreements if redistributing this
         package or utilizing the results based on this package.
         See the LICENSE file for more details."
     )
-    return(list(nodes = nodes, edges = edges))
+    return(subnetwork)
 }

R/utils_getSubnetworkFromIndra.R

@@ -68,9 +68,9 @@
 
 #' @importFrom httr GET status_code content
 #' @importFrom jsonlite fromJSON
-.get_incorrect_curation_count <- function(stmt_hash, api_key) {
+.get_incorrect_curation_count <- function(stmt_hash) {
     stmt_hash_char <- as.character(stmt_hash)
-    url <- paste0("https://db.indra.bio/curation/list/", stmt_hash_char, "?api_key=", api_key)
+    url <- paste0("https://db.indra.bio/curation/list/", stmt_hash_char)
 
     tryCatch({
         response <- GET(url)
@@ -96,21 +96,18 @@
 
 #' Call INDRA Cogex API and return response
 #' @param res response from INDRA
-#' @param interaction_types interaction types to filter by
+#' @param statement_types interaction types to filter by
 #' @param evidence_count_cutoff number of evidence to filter on for each paper
 #' @param sources_filter list of sources to filter by. Default is NULL, i.e. no filter
-#' @param filter_by_curation logical, whether to filter out statements that
-#' have been curated as incorrect in INDRA.  Default is FALSE.
-#' @param api_key string of INDRA API key for accessing curated statements.
 #' @return filtered list of INDRA statements
 #' @importFrom jsonlite fromJSON
 #' @keywords internal
 #' @noRd
-.filterIndraResponse <- function(res, interaction_types, evidence_count_cutoff, 
-                                 sources_filter = NULL, filter_by_curation = FALSE, api_key = "") {
-    if (!is.null(interaction_types)) {
+.filterIndraResponse <- function(res, statement_types, evidence_count_cutoff, 
+                                 sources_filter = NULL) {
+    if (!is.null(statement_types)) {
         res = Filter(
-            function(statement) statement$data$stmt_type %in% interaction_types, 
+            function(statement) statement$data$stmt_type %in% statement_types, 
             res)
     }
     if (!is.null(sources_filter)) {
@@ -123,16 +120,6 @@
             res
         )
     }
-    if (filter_by_curation) {
-        for (i in seq_along(res)) {
-            stmt_json <- fromJSON(res[[i]]$data$stmt_json)
-            stmt_hash <- stmt_json$matches_hash
-            incorrect_count <- .get_incorrect_curation_count(stmt_hash, api_key)
-            res[[i]]$data$evidence_count <- res[[i]]$data$evidence_count - incorrect_count
-            # Todo: Also subtract source_counts accordingly if requested
-            Sys.sleep(0.1)
-        }
-    }
     res = Filter(
         function(statement) statement$data$evidence_count >= evidence_count_cutoff, 
         res
@@ -373,6 +360,31 @@
     return(edges)
 }
 
+.filterByCuration = function(nodes, edges, evidence_count_cutoff, filter_by_curation) {
+    if (filter_by_curation) {
+        incorrect_counts <- numeric(nrow(edges))
+        for (i in seq_len(nrow(edges))) {
+            incorrect_counts[i] <- .get_incorrect_curation_count(edges$stmt_hash[i])
+            Sys.sleep(0.1)
+        }
+        edges$evidenceCount <- edges$evidenceCount - incorrect_counts
+        edges <- edges[edges$evidenceCount >= evidence_count_cutoff, ]
+        nodes <- nodes[nodes$id %in% c(edges$source, edges$target), ]
+    }
+    return(list(nodes = nodes, edges = edges))
+}
+
+.filterByPtmSite = function(nodes, edges, filter_by_ptm_site) {
+    if (filter_by_ptm_site && nrow(nodes[!is.na(nodes$Site), ]) > 0) {
+        ptm_overlap <- .calculatePTMOverlapAggregated(edges, nodes)
+        keep <- ptm_overlap[paste(edges$source, edges$target, edges$interaction, sep = "-")]
+        edges <- edges[!is.na(keep) & keep != "", ]
+        edges <- edges[!is.na(edges$site),]
+        nodes <- nodes[nodes$id %in% c(edges$source, edges$target), ]
+    }
+    return(list(nodes = nodes, edges = edges))
+}
+
 #' Construct correlation matrix from MSstats
 #' @param protein_level_data output of dataProcess
 #' @importFrom tidyr pivot_wider

R/visualizeNetworksWithHTML.R

@@ -78,8 +78,9 @@ getRelationshipProperties <- function() {
 #' @param edges Data frame with edge information including 'target' and 'site' columns
 #' @param nodes Data frame with node information including 'id' and 'Site' columns
 #' @return Vector of overlap descriptions for each unique edge (after consolidation)
+#' @keywords internal
 #' @noRd
-calculatePTMOverlapAggregated <- function(edges, nodes) {
+.calculatePTMOverlapAggregated <- function(edges, nodes) {
     if (nrow(edges) == 0) return(character(0))
 
     # Group edges by source-target-interaction to match consolidation logic
@@ -153,7 +154,7 @@ consolidateEdges <- function(edges, nodes = NULL) {
 
     # Calculate aggregated PTM overlap information if nodes are provided
     ptm_overlap_map <- if (!is.null(nodes)) {
-        calculatePTMOverlapAggregated(edges, nodes)
+        .calculatePTMOverlapAggregated(edges, nodes)
     } else {
         NULL
     }
@@ -1141,6 +1142,7 @@ createEdgeClickHandler <- function() {
 #' @param edges Data frame with edge information  
 #' @param filename Output HTML filename
 #' @param displayLabelType Type of label to display ("id" or "hgncName")
+#' @param nodeFontSize Font size for node labels (default: 12)
 #' @param ... Additional arguments passed to exportCytoscapeToHTML()
 #' @export
 #' @return Invisibly returns the file path of the created HTML file