diff --git a/DESCRIPTION b/DESCRIPTION
index 84fff48..e869e86 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -23,8 +23,7 @@ Imports:
httr,
jsonlite,
r2r,
- tidyr,
- MASS
+ tidyr
Suggests:
data.table,
BiocStyle,
diff --git a/NAMESPACE b/NAMESPACE
index b2e4376..e84ce42 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -3,11 +3,9 @@
export(annotateProteinInfoFromIndra)
export(exportNetworkToHTML)
export(generateCytoscapeConfig)
-export(getPathwaysFromIndra)
export(getSubnetworkFromIndra)
export(previewNetworkInBrowser)
export(visualizeNetworks)
-importFrom(MASS,fitdistr)
importFrom(RCy3,addAnnotationShape)
importFrom(RCy3,addAnnotationText)
importFrom(RCy3,createNetworkFromDataFrames)
@@ -20,7 +18,6 @@ importFrom(RCy3,mapVisualProperty)
importFrom(RCy3,setVisualStyle)
importFrom(grDevices,colorRamp)
importFrom(grDevices,rgb)
-importFrom(httr,GET)
importFrom(httr,POST)
importFrom(httr,add_headers)
importFrom(httr,content)
diff --git a/R/annotateProteinInfoFromIndra.R b/R/annotateProteinInfoFromIndra.R
index b4470b7..cffe254 100644
--- a/R/annotateProteinInfoFromIndra.R
+++ b/R/annotateProteinInfoFromIndra.R
@@ -55,18 +55,27 @@ annotateProteinInfoFromIndra <- function(df, proteinIdType) {
#' It can be either "Uniprot" or "Uniprot_Mnemonic".
#' @return A data frame with populated Uniprot IDs.
.populateUniprotIdsInDataFrame <- function(df, proteinIdType) {
+ if ("GlobalProtein" %in% colnames(df)) {
+ protein_ids = unique(as.character(df$GlobalProtein))
+ } else {
+ df$Protein = as.character(df$Protein)
+ df$GlobalProtein = ifelse(grepl("_[A-Z][0-9]", df$Protein),
+ gsub("_[A-Z][0-9].*", "", df$Protein, perl = TRUE),
+ df$Protein)
+ protein_ids = unique(df$GlobalProtein)
+ }
df$UniprotId <- NA
if (proteinIdType == "Uniprot") {
- df$UniprotId <- as.character(df$Protein)
+ df$UniprotId <- as.character(df$GlobalProtein)
}
if (proteinIdType == "Uniprot_Mnemonic") {
- mnemonicProteins <- df$Protein
+ mnemonicProteins <- protein_ids
if (length(mnemonicProteins) > 0) {
uniprotMapping <- .callGetUniprotIdsFromUniprotMnemonicIdsApi(as.list(mnemonicProteins))
for (mnemonicId in names(uniprotMapping)) {
if (!is.null(uniprotMapping[[mnemonicId]])) {
- df$UniprotId[df$Protein == mnemonicId] <- uniprotMapping[[mnemonicId]]
+ df$UniprotId[df$GlobalProtein == mnemonicId] <- uniprotMapping[[mnemonicId]]
}
}
}
diff --git a/R/getPathwaysFromIndra.R b/R/getPathwaysFromIndra.R
deleted file mode 100644
index 76c654e..0000000
--- a/R/getPathwaysFromIndra.R
+++ /dev/null
@@ -1,176 +0,0 @@
-#' Get pathways ranked on relevance from INDRA DB
-#'
-#' @importFrom httr GET content
-#' @importFrom MASS fitdistr
-#' @importFrom r2r hashmap keys query
-#'
-#' @param annotated_df output of \code{\link[MSstats]{groupComparison}} function's
-#' comparisionResult table, which contains a list of proteins and their
-#' corresponding p-values, logFCs, along with additional HGNC ID and HGNC
-#' name columns
-#' @param main_target A main target, e.g. main target of a drug or protein of
-#' particular interest
-#' @param target_type One of either 'Protein' or 'Drug'. Default is 'Protein'
-#'
-#' @return df of pathways
-#'
-#' @export
-#'
-#' @examples
-#' annotated_df <- data.table::fread(system.file(
-#' "extdata/groupComparisonModel.csv",
-#' package = "MSstatsBioNet"
-#' ))
-#' pathways <- getPathwaysFromIndra(annotated_df, "P05067")
-#' head(pathways)
-#'
-getPathwaysFromIndra <- function(annotated_df, main_target = 'MEN1_HUMAN', target_type = "Protein") {
- annotated_df$Protein <- as.character(annotated_df$Protein)
- log2fc_values <- annotated_df$log2FC
- fit <- fitdistr(log2fc_values, "normal")
- para <- fit$estimate
- # pnorm(0, mean = para[1], sd = para[2])
-
- # Fit a power-law-like curve
- b = -0.08
- m = -1.33
- # 1 - 10^(m*log10(1000)+b)
- # probability
-
- # Call INDRA
- if (target_type == "Protein") {
- main_target_row = annotated_df[annotated_df$Protein == main_target,]
- source_id = as.character(main_target_row$HgncId)
- namespace = "@HGNC"
- id_field = "HGNC"
- } else if (target_type == "Drug") {
- source_id = main_target
- namespace = ""
- id_field = "TEXT"
- } else {
- stop("Invalid target type.")
- }
- url = paste('https://db.indra.bio/statements/from_agents?subject=',
- source_id, namespace, sep = "")
- response <- GET(url)
- z = content(response)
-
- edgeToMetadataMapping <- hashmap()
-
- if (length(z$statements) == 0) {
- return(data.frame())
- }
-
- for (index in seq(1, length(z$statements))) {
- edge <- z$statements[[index]]
- if (edge$type == "Complex") {
- if (length(edge$members) == 2) {
- if (identical(edge$members[[1]]$db_refs[[id_field]], source_id)) {
- obj = edge$members[[2]]$db_refs$HGNC
- namespaces = names(edge$members[[2]]$db_refs)
- } else {
- obj = edge$members[[1]]$db_refs$HGNC
- namespaces = names(edge$members[[1]]$db_refs)
- }
- } else {
- namespaces = c()
- }
- } else if (edge$type == "Phosphorylation") {
- obj = edge$sub$db_refs$HGNC
- namespaces = names(edge$sub$db_refs)
- } else {
- obj = edge$obj$db_refs$HGNC
- namespaces = names(edge$obj$db_refs)
- }
-
- # Filter out edges with no HGNC ID or not in the dataset
- if (!("HGNC" %in% namespaces)) {
- next
- } else if (!(obj %in% annotated_df$HgncId)) {
- next
- }
-
- key <- paste(source_id, obj, sep = "_")
-
- if (key %in% keys(edgeToMetadataMapping)) {
- edgeToMetadataMapping[[key]]$data$evidence_count <-
- edgeToMetadataMapping[[key]]$data$evidence_count +
- z$evidence_counts[[index]]
- edgeToMetadataMapping[[key]]$data$stmt_type <- unique(c(
- edgeToMetadataMapping[[key]]$data$stmt_type,
- edge$type))
- } else {
- # edge <- MSstatsBioNet:::.addAdditionalMetadataToIndraEdge(edge, annotated_df)
- edgeToMetadataMapping[[key]] <- edge
- edgeToMetadataMapping[[key]]$data$evidence_count <-
- z$evidence_counts[[index]]
- edgeToMetadataMapping[[key]]$data$stmt_type <- c(edge$type)
- edgeToMetadataMapping[[key]]$source_id <- source_id
- edgeToMetadataMapping[[key]]$target_id <- obj
- edgeToMetadataMapping[[key]] <- .addAdditionalMetadataToIndraEdge(
- edgeToMetadataMapping[[key]], annotated_df
- )
- }
- }
-
- # Calculate probabilities
- for (key in keys(edgeToMetadataMapping)) {
- edgeToMetadataMapping[[key]]$data$stmt_type <-
- paste(unique(edgeToMetadataMapping[[key]]$data$stmt_type),
- collapse = ", ")
- prob_logFC = 0
- logFC = annotated_df[which(annotated_df$HgncId == edgeToMetadataMapping[[key]]$target_id),]
- logFC = logFC$log2FC[[1]]
- if (logFC > para[1]) {
- prob_logFC = 1 - pnorm(logFC, mean = para[1], sd = para[2])
- } else {
- prob_logFC = pnorm(logFC, mean = para[1], sd = para[2])
- }
- evidence_prob = 10^(m*log10(edgeToMetadataMapping[[key]]$data$evidence_count)+b)
- edgeToMetadataMapping[[key]]$data$total_prob = 1 - ((1 - prob_logFC) * (1 - evidence_prob))
- edgeToMetadataMapping[[key]]$data$logFC = logFC
- }
-
- # Construct DF and sort
- edges <- data.frame(
- source = vapply(keys(edgeToMetadataMapping), function(x) {
- query(edgeToMetadataMapping, x)$source_uniprot_id
- }, ""),
- target = vapply(keys(edgeToMetadataMapping), function(x) {
- query(edgeToMetadataMapping, x)$target_uniprot_id
- }, ""),
- interaction = vapply(keys(edgeToMetadataMapping), function(x) {
- query(edgeToMetadataMapping, x)$data$stmt_type
- }, ""),
- evidenceCount = vapply(keys(edgeToMetadataMapping), function(x) {
- query(edgeToMetadataMapping, x)$data$evidence_count
- }, 1),
- logFC = vapply(keys(edgeToMetadataMapping), function(x) {
- query(edgeToMetadataMapping, x)$data$logFC
- }, 1),
- prob = vapply(keys(edgeToMetadataMapping), function(x) {
- query(edgeToMetadataMapping, x)$data$total_prob
- }, 1),
- evidenceLink = vapply(keys(edgeToMetadataMapping), function(x) {
- query(edgeToMetadataMapping, x)$evidence_list
- }, ""),
- stringsAsFactors = FALSE
- )
-
- nodes <- .constructNodesDataFrame(
- annotated_df, edges
- )
- if (!(main_target %in% nodes$id)) {
- # add a row with the main target
- nodes <- rbind(
- nodes,
- data.frame(
- id = main_target,
- logFC = 0,
- pvalue = 0,
- hgncName = main_target
- )
- )
- }
- return(list(nodes = nodes, edges = edges))
-}
\ No newline at end of file
diff --git a/R/utils_getSubnetworkFromIndra.R b/R/utils_getSubnetworkFromIndra.R
index e514d52..f3b8bce 100644
--- a/R/utils_getSubnetworkFromIndra.R
+++ b/R/utils_getSubnetworkFromIndra.R
@@ -10,7 +10,8 @@
if (!"HgncId" %in% colnames(input)) {
stop("Invalid Input Error: Input must contain a column named 'HgncId'.")
}
- num_proteins = nrow(input) + ifelse(!is.null(force_include_other), length(force_include_other), 0)
+ num_proteins = length(unique(input$HgncId)) +
+ ifelse(!is.null(force_include_other), length(force_include_other), 0)
if (num_proteins >= 400) {
stop("Invalid Input Error: INDRA query must contain less than 400 proteins. Consider lowering your p-value cutoff")
}
@@ -41,6 +42,7 @@
indraCogexUrl <-
"https://discovery.indra.bio/api/indra_subnetwork_relations"
+ hgncIds = unique(hgncIds)
groundings <- lapply(hgncIds, function(x) list("HGNC", x))
if (!is.null(force_include_other)) {
groundings <- c(groundings, lapply(force_include_other, function(x) {
@@ -105,6 +107,8 @@
#' @keywords internal
#' @noRd
.filterGetSubnetworkFromIndraInput <- function(input, pvalueCutoff, logfc_cutoff, force_include_proteins) {
+ input$Protein <- as.character(input$Protein)
+
# Extract exempt proteins before any filtering
exempt_proteins <- NULL
if (!is.null(force_include_proteins)) {
@@ -140,7 +144,16 @@
input <- combined_input[!duplicated(combined_input$Protein), ]
}
- input$Protein <- as.character(input$Protein)
+ # Handle PTMs in Protein column
+ input$Site = ifelse(grepl("_[A-Z][0-9]", input$Protein),
+ gsub("^_", "",
+ gsub("^[^_]*_|_(?![A-Z][0-9])[^_]*", "", input$Protein, perl = TRUE)
+ ),
+ NA_character_
+ )
+ if ("GlobalProtein" %in% colnames(input)) {
+ input$Protein = input$GlobalProtein
+ }
return(input)
}
#' Add additional metadata to an edge
@@ -159,17 +172,19 @@
# Convert back to uniprot IDs
matched_rows_source <- input[which(input$HgncId == edge$source_id), ]
- if (nrow(matched_rows_source) != 1) {
+ uniprot_ids_source <- unique(matched_rows_source$Protein)
+ if (length(uniprot_ids_source) != 1) {
edge$source_uniprot_id <- edge$source_name
} else {
- edge$source_uniprot_id <- matched_rows_source$Protein
+ edge$source_uniprot_id <- uniprot_ids_source
}
matched_rows_target <- input[which(input$HgncId == edge$target_id), ]
- if (nrow(matched_rows_target) != 1) {
+ uniprot_ids_target = unique(matched_rows_target$Protein)
+ if (length(uniprot_ids_target) != 1) {
edge$target_uniprot_id <- edge$target_name
} else {
- edge$target_uniprot_id <- matched_rows_target$Protein
+ edge$target_uniprot_id <- uniprot_ids_target
}
return(edge)
@@ -179,6 +194,7 @@
#' Collapse duplicate INDRA statements into a mapping of edge to metadata
#' @param res INDRA response
#' @param input filtered groupComparison result
+#' @importFrom jsonlite fromJSON
#' @importFrom r2r hashmap keys
#' @return processed edge to metadata mapping
#' @keywords internal
@@ -188,6 +204,13 @@
for (edge in res) {
key <- paste(edge$source_id, edge$target_id, edge$data$stmt_type, sep = "_")
+ json_object <- fromJSON(edge$data$stmt_json)
+ if (!is.null(json_object$residue) && !is.null(json_object$position)) {
+ edge$site = paste0(json_object$residue, json_object$position)
+ key <- paste(key, edge$site, sep = "_")
+ } else {
+ edge$site = NA_character_
+ }
if (key %in% keys(edgeToMetadataMapping)) {
edgeToMetadataMapping[[key]]$data$evidence_count <-
edgeToMetadataMapping[[key]]$data$evidence_count +
@@ -209,6 +232,7 @@
#' @param input filtered groupComparison result
#' @param protein_level_data output of dataProcess
#' @importFrom r2r query keys
+#' @importFrom jsonlite fromJSON
#' @return edge data.frame
#' @keywords internal
#' @noRd
@@ -236,6 +260,9 @@
sourceCounts = vapply(keys(res), function(x) {
query(res, x)$data$source_counts
}, ""),
+ site = vapply(keys(res), function(x) {
+ query(res, x)$site
+ }, ""),
stringsAsFactors = FALSE
)
# add correlation - maybe create a separate function
@@ -262,24 +289,11 @@
#' @keywords internal
#' @noRd
.constructNodesDataFrame <- function(input, edges) {
- # Get unique nodes from edges
- node_ids <- unique(c(edges$source, edges$target))
+ nodes = input[, c("Protein", "log2FC", "adj.pvalue", "HgncName", "Site")]
+ colnames(nodes) = c("id", "logFC", "adj.pvalue", "hgncName", "Site")
- # Create base nodes dataframe
- nodes <- data.frame(
- id = node_ids,
- stringsAsFactors = FALSE
- )
-
- # Add attributes from input where available
- nodes$logFC <- input$log2FC[match(nodes$id, input$Protein)]
- nodes$adj.pvalue <- input$adj.pvalue[match(nodes$id, input$Protein)]
- nodes$hgncName <- if ("HgncName" %in% colnames(input) && is.character(input$HgncName)) {
- hgnc_value <- input$HgncName[match(nodes$id, input$Protein)]
- ifelse(is.na(hgnc_value), nodes$id, hgnc_value)
- } else {
- nodes$id
- }
+ nodes = nodes[nodes$id %in% c(edges$source, edges$target), ]
+ nodes$hgncName = ifelse(is.na(nodes$hgncName), nodes$id, nodes$hgncName)
return(nodes)
}
diff --git a/R/visualizeNetworksWithHTML.R b/R/visualizeNetworksWithHTML.R
index cd4772b..d10eb33 100644
--- a/R/visualizeNetworksWithHTML.R
+++ b/R/visualizeNetworksWithHTML.R
@@ -1,8 +1,3 @@
-# =============================================================================
-# CYTOSCAPE VISUALIZATION PACKAGE FUNCTIONS
-# These functions should go in your separate package
-# =============================================================================
-
#' Helper function to map logFC values to colors
#' @param logFC_values Numeric vector of log fold change values
#' @importFrom grDevices colorRamp rgb
@@ -79,8 +74,75 @@ getRelationshipProperties <- function() {
)
}
+#' Calculate PTM site overlap between edge targets and nodes with aggregation
+#' @param edges Data frame with edge information including 'target' and 'site' columns
+#' @param nodes Data frame with node information including 'id' and 'Site' columns
+#' @return Vector of overlap descriptions for each unique edge (after consolidation)
+#' @noRd
+calculatePTMOverlapAggregated <- function(edges, nodes) {
+ if (nrow(edges) == 0) return(character(0))
+
+ # Group edges by source-target-interaction to match consolidation logic
+ edges$edge_key <- paste(edges$source, edges$target, edges$interaction, sep = "-")
+ unique_edges <- unique(edges$edge_key)
+
+ overlap_info <- character(length(unique_edges))
+ names(overlap_info) <- unique_edges
+
+ for (edge_key in unique_edges) {
+ # Get all edges with this source-target-interaction combination
+ matching_edges <- edges[edges$edge_key == edge_key, ]
+ all_overlap_sites <- c()
+
+ # Process each matching edge to find PTM overlaps
+ for (i in 1:nrow(matching_edges)) {
+ edge <- matching_edges[i, ]
+
+ # Check if edge has target and site information
+ if (!is.na(edge$target) && "site" %in% names(edge) && !is.na(edge$site)) {
+ # Find matching nodes with the same target ID
+ target_nodes <- nodes[nodes$id == edge$target, ]
+
+ if (nrow(target_nodes) > 0 && "Site" %in% names(target_nodes)) {
+ edge_sites <- trimws(unlist(strsplit(as.character(edge$site), "[,;|]")))
+
+ # Check each target node row for site matches
+ for (j in 1:nrow(target_nodes)) {
+ if (!is.na(target_nodes$Site[j])) {
+ node_sites <- trimws(unlist(strsplit(as.character(target_nodes$Site[j]), "_")))
+
+ # Find overlapping sites for this edge-node combination
+ overlap_sites <- intersect(edge_sites, node_sites)
+ overlap_sites <- overlap_sites[overlap_sites != "" & !is.na(overlap_sites)]
+
+ # Add to the aggregate list
+ all_overlap_sites <- c(all_overlap_sites, overlap_sites)
+ }
+ }
+ }
+ }
+ }
+
+ # Remove duplicates and create tooltip text for this consolidated edge
+ unique_overlap_sites <- unique(all_overlap_sites)
+ unique_overlap_sites <- unique_overlap_sites[unique_overlap_sites != "" & !is.na(unique_overlap_sites)]
+
+ if (length(unique_overlap_sites) > 0) {
+ if (length(unique_overlap_sites) == 1) {
+ overlap_info[edge_key] <- paste0("Overlapping PTM site: ", unique_overlap_sites[1])
+ } else {
+ overlap_info[edge_key] <- paste0("Overlapping PTM sites: ", paste(unique_overlap_sites, collapse = ", "))
+ }
+ } else {
+ overlap_info[edge_key] <- ""
+ }
+ }
+
+ return(overlap_info)
+}
+
# Consolidate bidirectional edges based on relationship type
-consolidateEdges <- function(edges) {
+consolidateEdges <- function(edges, nodes = NULL) {
if (nrow(edges) == 0) return(edges)
required_cols <- c("source", "target", "interaction")
@@ -89,6 +151,13 @@ consolidateEdges <- function(edges) {
stop("Missing required columns: ", paste(missing_cols, collapse = ", "))
}
+ # Calculate aggregated PTM overlap information if nodes are provided
+ ptm_overlap_map <- if (!is.null(nodes)) {
+ calculatePTMOverlapAggregated(edges, nodes)
+ } else {
+ NULL
+ }
+
relationship_props <- getRelationshipProperties()
consolidated_edges <- list()
processed_pairs <- c()
@@ -96,7 +165,6 @@ consolidateEdges <- function(edges) {
for (i in 1:nrow(edges)) {
edge <- edges[i, ]
pair_key <- paste(sort(c(edge$source, edge$target)), edge$interaction, collapse = "-")
- reverse_key <- paste(sort(c(edge$source, edge$target), decreasing = TRUE), edge$interaction, sep = "-")
# Skip if we've already processed this pair
if (pair_key %in% processed_pairs) next
@@ -118,6 +186,14 @@ consolidateEdges <- function(edges) {
consolidation_type <- relationship_props[[category]]$consolidate
+ # Get PTM overlap info for this edge combination
+ edge_key <- paste(edge$source, edge$target, edge$interaction, sep = "-")
+ ptm_overlap_text <- if (!is.null(ptm_overlap_map) && edge_key %in% names(ptm_overlap_map)) {
+ ptm_overlap_map[[edge_key]]
+ } else {
+ ""
+ }
+
if (nrow(reverse_edges) > 0 && consolidation_type %in% c("undirected", "bidirectional")) {
# Create consolidated edge
if (consolidation_type == "undirected") {
@@ -128,6 +204,7 @@ consolidateEdges <- function(edges) {
interaction = edge$interaction,
edge_type = "undirected",
category = category,
+ ptm_overlap = ptm_overlap_text,
stringsAsFactors = FALSE
)
} else {
@@ -138,6 +215,7 @@ consolidateEdges <- function(edges) {
interaction = paste(edge$interaction, "(bidirectional)"),
edge_type = "bidirectional",
category = category,
+ ptm_overlap = ptm_overlap_text,
stringsAsFactors = FALSE
)
}
@@ -148,8 +226,8 @@ consolidateEdges <- function(edges) {
consolidated_edge[[col]] <- edge[[col]]
}
- edge_key <- paste(edge$source, edge$target, consolidated_edge$interaction, sep = "-")
- consolidated_edges[[edge_key]] <- consolidated_edge
+ edge_key_final <- paste(edge$source, edge$target, consolidated_edge$interaction, sep = "-")
+ consolidated_edges[[edge_key_final]] <- consolidated_edge
# Mark both directions as processed
processed_pairs <- c(processed_pairs, pair_key)
@@ -159,9 +237,10 @@ consolidateEdges <- function(edges) {
directed_edge <- edge
directed_edge$edge_type <- "directed"
directed_edge$category <- category
+ directed_edge$ptm_overlap = ptm_overlap_text
- edge_key <- paste(edge$source, edge$target, edge$interaction, sep = "-")
- consolidated_edges[[edge_key]] <- directed_edge
+ edge_key_final <- paste(edge$source, edge$target, edge$interaction, sep = "-")
+ consolidated_edges[[edge_key_final]] <- directed_edge
}
}
@@ -242,11 +321,11 @@ createNodeElements <- function(nodes, displayLabelType = "id") {
})
}
-createEdgeElements <- function(edges) {
+createEdgeElements <- function(edges, nodes = NULL) {
if (nrow(edges) == 0) return(list())
# First consolidate edges
- consolidated_edges <- consolidateEdges(edges)
+ consolidated_edges <- consolidateEdges(edges, nodes)
edge_elements <- list()
@@ -262,7 +341,10 @@ createEdgeElements <- function(edges) {
evidence_link <- ifelse(is.na(evidence_link) | evidence_link == "NA", "", evidence_link)
evidence_link <- escape_js_string(evidence_link)
- # Create edge data with styling information
+ # Escape quotes in tooltip text for JavaScript safety
+ tooltip_text <- gsub("'", "\\\\'", row$ptm_overlap)
+
+ # Create edge data with styling information and PTM overlap tooltip
edge_data <- paste0("{ data: { source: '", row$source,
"', target: '", row$target,
"', id: '", edge_key,
@@ -273,7 +355,8 @@ createEdgeElements <- function(edges) {
"', color: '", style$color,
"', line_style: '", style$style,
"', arrow_shape: '", style$arrow,
- "', width: ", style$width, " } }")
+ "', width: ", style$width,
+ ", tooltip: '", tooltip_text, "' } }")
edge_elements[[edge_key]] <- edge_data
}
@@ -319,7 +402,7 @@ generateCytoscapeConfig <- function(nodes, edges,
# Create elements
node_elements <- createNodeElements(nodes, display_label_type)
- edge_elements <- createEdgeElements(edges)
+ edge_elements <- createEdgeElements(edges, nodes)
# Default layout options
default_layout <- list(
@@ -342,7 +425,7 @@ generateCytoscapeConfig <- function(nodes, edges,
}
}
- # Define the style configuration (same as before)
+ # Define the style configuration
style_config <- list(
list(
selector = "node",
@@ -470,6 +553,48 @@ generateJavaScriptCode <- function(config) {
layout: ", layout_js, "
});
+ // Create tooltip element
+ var tooltip = document.createElement('div');
+ tooltip.style.cssText = `
+ position: absolute;
+ background-color: rgba(0, 0, 0, 0.9);
+ color: white;
+ padding: 8px 12px;
+ border-radius: 4px;
+ font-size: 12px;
+ font-family: Arial, sans-serif;
+ white-space: nowrap;
+ pointer-events: none;
+ z-index: 9999;
+ box-shadow: 0 2px 8px rgba(0, 0, 0, 0.3);
+ display: none;
+ max-width: 300px;
+ word-wrap: break-word;
+ white-space: pre-wrap;
+ `;
+ document.body.appendChild(tooltip);
+
+ // Only show tooltip if there's actual PTM overlap information
+ cy.on('mouseover', 'edge', function(evt) {
+ var edge = evt.target;
+ var tooltipText = edge.data('tooltip');
+ if (tooltipText && tooltipText.trim() !== '' && tooltipText.trim() !== 'No overlapping PTM sites found') {
+ tooltip.innerHTML = tooltipText;
+ tooltip.style.display = 'block';
+ }
+ });
+
+ cy.on('mousemove', 'edge', function(evt) {
+ if (tooltip.style.display === 'block') {
+ tooltip.style.left = evt.originalEvent.pageX + 10 + 'px';
+ tooltip.style.top = evt.originalEvent.pageY - 30 + 'px';
+ }
+ });
+
+ cy.on('mouseout', 'edge', function(evt) {
+ tooltip.style.display = 'none';
+ });
+
", event_handlers_js)
return(js_code)
@@ -665,7 +790,7 @@ exportCytoscapeToHTML <- function(config,
}
#'
- , config$container_id, ' {
+ , config$container_id, ' {
width: ', width, ';
height: ', height, ';
border: 1px solid #ddd;
@@ -768,6 +893,7 @@ exportCytoscapeToHTML <- function(config,
Instructions:
Click and drag to pan the network
| Use mouse wheel to zoom in/out
+ | Hover over edges to see PTM site overlap information
| Click on nodes or edges to select them
', if(include_controls) '| Use the buttons above for common navigation actions' else '', '
@@ -846,6 +972,10 @@ exportCytoscapeToHTML <- function(config,
Complex
+
+ PTM Site Info:
+ Hover over edges to see overlapping PTM sites between the edge target and node data
+
`;
legendDiv.innerHTML = legendHTML;
@@ -934,7 +1064,7 @@ exportNetworkToHTML <- function(nodes, edges,
previewNetworkInBrowser <- function(nodes, edges,
displayLabelType = "id",
...) {
-
+
# Generate configuration
config <- generateCytoscapeConfig(nodes, edges, display_label_type = displayLabelType)
diff --git a/inst/extdata/garrido-2024.csv b/inst/extdata/garrido-2024.csv
new file mode 100644
index 0000000..27e174a
--- /dev/null
+++ b/inst/extdata/garrido-2024.csv
@@ -0,0 +1,23 @@
+"Protein","Label","log2FC","SE","DF","pvalue","adj.pvalue","issue"
+"P00533_S1039_S1042","t0 vs t1",-0.32003632837107,0.149943118877325,9,0.0615809510693193,0.280245896469974,NA
+"P00533_S1064","t0 vs t1",0.356653113576368,0.0891534693290476,9,0.00310836383475933,0.0686359841548604,NA
+"P00533_S991_S995","t0 vs t1",-0.122903662291889,0.10858297785538,9,0.286937654558352,0.579073736150236,NA
+"P00533_T693","t0 vs t1",-0.0233443995453799,0.177244588005518,9,0.898113182376962,0.96083633837953,NA
+"P00533_T693_S695","t0 vs t1",-0.165995656836019,0.150007538230776,9,0.297173768898177,0.588091082037082,NA
+"P00533_Y1110","t0 vs t1",0.210632429258209,0.0927903082311584,9,0.0493644341170289,0.252589138618561,NA
+"P28482_T185","t0 vs t1",0.634378738770068,0.182300631518675,9,0.00694020442220732,0.0999180588053667,NA
+"P28482_T185_T190","t0 vs t1",0.890196707480898,0.165905887251245,9,0.000452927502292055,0.025493055208194,NA
+"P28482_T185_Y187","t0 vs t1",0.77570963567424,0.200315650539845,9,0.0037747061619258,0.0748119597579435,NA
+"P28482_Y187","t0 vs t1",1.13798944558515,0.182042606245047,9,0.000149415500661671,0.0138506354758592,NA
+"P28482_Y187_T190","t0 vs t1",0.0839586116733722,0.114930246570557,9,0.483655330294773,0.739194565873987,NA
+"Q13480_S266","t0 vs t1",0.147775796440306,0.106174162166453,9,0.19741006608831,0.483926492931687,NA
+"Q13480_S418_S419","t0 vs t1",0.281122057791004,0.146958322587612,9,0.0880466739278739,0.328235537559088,NA
+"Q13480_S419","t0 vs t1",0.320048191455162,0.114579108916761,9,0.0209422984694783,0.166454473266352,NA
+"Q13480_S500_T503","t0 vs t1",-0.450413295801827,0.0980746173074297,9,0.00130426958782228,0.0449838460341635,NA
+"Q13480_S634_S637","t0 vs t1",-0.616943295142088,0.129109930148088,9,0.00100346794179792,0.0394342642255919,NA
+"Q13480_S650_S651_Y659","t0 vs t1",1.42523331569634,0.173080560297211,9,1.75594329256568e-05,0.00503640093293397,NA
+"Q13480_S651_Y659","t0 vs t1",1.34708073620391,0.292405749303692,9,0.00127794033828973,0.0447363035684408,NA
+"Q13480_T387_S389","t0 vs t1",-0.494395484918809,0.116320309251806,9,0.0021415442309728,0.0576989604963562,NA
+"Q13480_Y265_S266","t0 vs t1",0.438767055281507,0.111182087237631,9,0.00337315512046587,0.0713587661347963,NA
+"Q13480_Y627_T638","t0 vs t1",2.58471343768464,0.276077760810949,9,6.17542514318978e-06,0.0032951498209992,NA
+"Q13480_Y659","t0 vs t1",2.93366028129369,0.180356840960513,9,5.57098141249871e-08,0.000346487188950357,NA
diff --git a/man/getPathwaysFromIndra.Rd b/man/getPathwaysFromIndra.Rd
deleted file mode 100644
index 9f37110..0000000
--- a/man/getPathwaysFromIndra.Rd
+++ /dev/null
@@ -1,38 +0,0 @@
-% Generated by roxygen2: do not edit by hand
-% Please edit documentation in R/getPathwaysFromIndra.R
-\name{getPathwaysFromIndra}
-\alias{getPathwaysFromIndra}
-\title{Get pathways ranked on relevance from INDRA DB}
-\usage{
-getPathwaysFromIndra(
- annotated_df,
- main_target = "MEN1_HUMAN",
- target_type = "Protein"
-)
-}
-\arguments{
-\item{annotated_df}{output of \code{\link[MSstats]{groupComparison}} function's
-comparisionResult table, which contains a list of proteins and their
-corresponding p-values, logFCs, along with additional HGNC ID and HGNC
-name columns}
-
-\item{main_target}{A main target, e.g. main target of a drug or protein of
-particular interest}
-
-\item{target_type}{One of either 'Protein' or 'Drug'. Default is 'Protein'}
-}
-\value{
-df of pathways
-}
-\description{
-Get pathways ranked on relevance from INDRA DB
-}
-\examples{
-annotated_df <- data.table::fread(system.file(
- "extdata/groupComparisonModel.csv",
- package = "MSstatsBioNet"
-))
-pathways <- getPathwaysFromIndra(annotated_df, "P05067")
-head(pathways)
-
-}
diff --git a/vignettes/MSstatsBioNet.Rmd b/vignettes/MSstatsBioNet.Rmd
index 8c9c1d5..d3fd498 100644
--- a/vignettes/MSstatsBioNet.Rmd
+++ b/vignettes/MSstatsBioNet.Rmd
@@ -106,7 +106,11 @@ subnetwork of proteins from the INDRA database based on differential abundance
analysis results.
```{r}
-subnetwork <- getSubnetworkFromIndra(annotated_df, pvalueCutoff = 0.05)
+subnetwork <- getSubnetworkFromIndra(
+ annotated_df,
+ pvalueCutoff = 0.05,
+ statement_types = c("Complex", "IncreaseAmount", "DecreaseAmount", "Inhibition", "Activation", "Phosphorylation")
+)
head(subnetwork$nodes)
head(subnetwork$edges)
```
@@ -116,16 +120,18 @@ This package is distributed under the [Artistic-2.0](https://opensource.org/lice
## Visualize Networks
-The function `visualizeNetworks` then takes the output of
+The function `previewNetworkInBrowser` then takes the output of
`getSubnetworkFromIndra` and visualizes the subnetwork. The function requires
-Cytoscape desktop to be open for the visualization to work.
+an internet browser to view the subnetwork
```{r}
-visualizeNetworks(subnetwork$nodes, subnetwork$edges)
+previewNetworkInBrowser(subnetwork$nodes, subnetwork$edges, displayLabelType = "hgncName")
```
-In the network diagram displayed in Cytoscape, you should see two arrows
-connecting two nodes, P16050 and P84243. These arrows represent the
-interactions between these two proteins, notably activation and phosphorylation.
+
+In the network diagram displayed using CytoscapeJS, you should see three arrows
+connecting two nodes, APP and APOD. These arrows represent the
+interactions between these two proteins, notably Activation, IncreaseAmount
+(e.g. via transcriptional regulation), and forming a protein complex.
# Session info
diff --git a/vignettes/PTM-Analysis.Rmd b/vignettes/PTM-Analysis.Rmd
new file mode 100644
index 0000000..67bd7cc
--- /dev/null
+++ b/vignettes/PTM-Analysis.Rmd
@@ -0,0 +1,83 @@
+---
+title: "PTM Analysis"
+author: "Anthony Wu"
+package: MSstatsBioNet
+output: BiocStyle::html_document
+vignette: >
+ %\VignetteIndexEntry{MSstatsBioNet: PTM Analysis}
+ %\VignetteEngine{knitr::rmarkdown}
+ %\VignetteEncoding{UTF-8}
+---
+
+```{r, include = FALSE}
+knitr::opts_chunk$set(
+ collapse = TRUE,
+ comment = "#>"
+)
+```
+
+
+# Installation
+
+Run this code below to install MSstatsBioNet from bioconductor
+
+```{r eval = FALSE}
+if (!require("BiocManager", quietly = TRUE)) {
+ install.packages("BiocManager")
+}
+
+BiocManager::install("MSstatsBioNet")
+```
+
+## Dataset
+
+We will be taking a subset of the dataset found in this [paper](https://www.biorxiv.org/content/10.1101/2024.10.21.619348v1). The table is the output of the MSstatsPTM function `groupComparisonPTM` (filtered down to the columns that are actually needed)
+
+```{r}
+input = data.table::fread(system.file(
+ "extdata/garrido-2024.csv",
+ package = "MSstatsBioNet"
+))
+head(input)
+```
+
+## ID Conversion
+
+First, we need to convert the group comparison results to a format that can be
+processed by INDRA. We can use the `annotateProteinInfoFromIndra` function
+to obtain these mappings.
+
+In the below example, we convert uniprot IDs to their corresponding Hgnc IDs. We
+can also extract other information, such as hgnc gene name and protein function.
+
+```{r}
+library(MSstatsBioNet)
+annotated_df = annotateProteinInfoFromIndra(input, "Uniprot")
+head(annotated_df)
+```
+
+## Subnetwork Query
+
+The package provides a function `getSubnetworkFromIndra` that retrieves a
+subnetwork of proteins from the INDRA database based on differential abundance
+analysis results. This function may help finding off target subnetworks.
+
+```{r}
+subnetwork <- getSubnetworkFromIndra(annotated_df, pvalueCutoff = 0.05, statement_types = c("Phosphorylation"), logfc_cutoff = 1, force_include_proteins = c("P00533_Y1110"))
+head(subnetwork$nodes)
+head(subnetwork$edges)
+```
+
+## Network Visualization
+
+Visualize the subnetwork on your browser
+
+```{r}
+previewNetworkInBrowser(subnetwork$nodes, subnetwork$edges, displayLabelType = "hgncName")
+```
+
+# Session info
+
+```{r}
+sessionInfo()
+```