clean up

mpc-bioinformatics · Mar 25, 2024 · ba6131f · ba6131f
1 parent b1c014f
commit ba6131f
Show file tree

Hide file tree

Showing 13 changed files with 22 additions and 162 deletions.
diff --git a/R/generate_graphs_from_FASTA.R b/R/generate_graphs_from_FASTA.R
@@ -1,11 +1,9 @@
-### TODO: Matrix package wird auf jeden Fall benötigt
-
 #' Generate graphs from a FASTA file
 #'
 #' @param fasta fasta file, already read into R by seqinr::read.fasta
 #' @param collapse_protein_nodes collapse protein nodes?
 #' @param collapse_peptide_nodes collapse peptide nodes?
-#' @param result_path path whereresults are saved. If NULL, results are not saved
+#' @param result_path path where results are saved. If NULL, results are not saved
 #' @param suffix suffix for saving results
 #' @param save_intermediate Save intermediate results?
 #' @param ... additional arguments to bppg::digest_fasta()
@@ -28,7 +26,7 @@ generate_graphs_from_FASTA <- function(fasta, collapse_protein_nodes = TRUE,
                                        ...) {
 
   message("Digesting FASTA file...")
-  digested_proteins <- bppg::digest_fasta(fasta, ...)#, ...)
+  digested_proteins <- bppg::digest_fasta(fasta, ...)
   message("Generating edgelist ...")
   edgelist <- bppg::generate_edgelist(digested_proteins, prot_origin = prot_origin)
   if (save_intermediate) {
@@ -72,9 +70,3 @@ generate_graphs_from_FASTA <- function(fasta, collapse_protein_nodes = TRUE,
 
 
 
-###
-
-
-###
-
-
diff --git a/R/generate_graphs_from_quantdata.R b/R/generate_graphs_from_quantdata.R
@@ -1,5 +1,3 @@
-
-
 #' Generate graphs from peptide ratio table, using an edgelist calculated on the fasta file
 #'
 #' @param peptide_ratios table with peptide ratios
@@ -10,8 +8,6 @@
 #' @export
 #'
 #' @examples
-#' ### TODO: Einstellbar, ob Peptid-Knoten auch gemergt werden sollen (dann mit geom. Mittel als peptid-ratio).
-#' ####      Das funktioniert noch nicht!!!
 generate_quant_graphs <- function(peptide_ratios, id_cols = 1, fasta_edgelist, outpath = NULL, seq_column = "Sequence",
                                   collapse_protein_nodes = TRUE, collapse_peptide_nodes = FALSE, suffix = "") {
 
@@ -30,14 +26,9 @@ generate_quant_graphs <- function(peptide_ratios, id_cols = 1, fasta_edgelist, o
   colnames_split <- limma::strsplit2(colnames(peptide_ratios), "_")
   comparisons <- paste(colnames_split[,2], colnames_split[,3], sep = "_")
 
-
-  # graphs <- list()
   subgraphs <- list()
-  #i = 3
-
-  ### TODO: progress bar!
-  for (i in 1:ncol(peptide_ratios)) { #
 
+  for (i in 1:ncol(peptide_ratios)) {
     comparison <- comparisons[i]
 
     fc <- peptide_ratios[,i]
@@ -63,8 +54,6 @@ generate_quant_graphs <- function(peptide_ratios, id_cols = 1, fasta_edgelist, o
       G[[j]] <- igraph::set_vertex_attr(graph = G[[j]], name = "pep_ratio",
                                         index = igraph::V(G[[j]])[!igraph::V(G[[j]])$type],
                                         value = edgelist_coll$pep_ratio[match(igraph::V(G[[j]])$name[!igraph::V(G[[j]])$type], edgelist_coll$peptide)])
-
-      #peptide_ratios[match(igraph::V(G[[j]])$name[!igraph::V(G[[j]])$type], id[, seq_column]), i]
     }
 
     subgraphs[[i]] <- G
@@ -77,9 +66,6 @@ generate_quant_graphs <- function(peptide_ratios, id_cols = 1, fasta_edgelist, o
 }
 
 
-
-### TODO: save end and intermediate results
-
 #' Generate graphs from quantitative peptide-level data
 #'
 #' @param D data set with peptide sequence as first column and peptide intensities in subsequent columns
@@ -95,7 +81,7 @@ generate_quant_graphs <- function(peptide_ratios, id_cols = 1, fasta_edgelist, o
 #' @return list of list of graphs
 #' @export
 #'
-#' @examples # TODO
+#' @examples
 generate_graphs_from_quant_data <- function(D, fasta, outpath = NULL, normalize = FALSE,
                                             missed_cleavages = 2, min_aa = 6, max_aa = 50,
                                             id_columns = 1, seq_column = "Sequence",
@@ -115,17 +101,11 @@ generate_graphs_from_quant_data <- function(D, fasta, outpath = NULL, normalize
 
 
   # remove peptides outside the desired length range
-  # TODO: remove also peptides with too many missed cleavages
   D <- D[nchar(D[, seq_column]) >= min_aa & nchar(D[, seq_column]) <= max_aa,]
 
 
   #normalize Intensities
   intensities <- D[,-id_columns]
-  # TODO: auch Median, Quantilsnormalisierung und LFQ-Normalisierung von MaxQuant erlauben?
-  if (normalize) {
-    #intensities <- 2^limma::normalizeBetweenArrays(log2(intensities), method = "cyclicloess")
-  }
-
 
   ### aggregate replicates by calculating the mean
   group <- factor(limma::strsplit2(colnames(intensities), "_")[,1])

diff --git a/R/helpers-add_graph_attributes.R b/R/helpers-add_graph_attributes.R
@@ -1,6 +1,3 @@
-
-
-
 #' Adds vertex attributes with uniqueness of peptides and number of unique peptides
 #' for proteins
 #'

diff --git a/R/helpers-assign_protein_accessions.R b/R/helpers-assign_protein_accessions.R
@@ -1,4 +1,3 @@
-
 #' Assign protein accessions to a list of peptides, depending on a FASTA file.
 #'
 #' @param sequence vector of peptide sequences
@@ -43,28 +42,3 @@ assign_protein_accessions <- function(sequence, fasta_vec) {
   return(unlist(assigned_proteins))
 }
 
-### TODO: wird abgelöst durch Grapherstellung über die Edgelist (Studienprojekt WS 22/23)
-
-# sequence, fasta_vec
-
-# protein_accessions <- names(fasta_vec)
-#
-# ### 1000 -> 12min
-#
-# x <- pblapply(sequence, function(x) {
-#   ind <- grep(paste0("(?:^M|K|R|^)(", x, ")(?=[^P]|$)"), fasta_vec, perl = TRUE)
-#   ### TODO: passt noch nicht ganz 100%ig, weil
-#   #proteins <- protein_accessions[ind]
-#   #proteins <- BBmisc::collapse(proteins, sep = "/")
-#   return(ind)
-# }
-# )
-# #
-# #
-# #
-# # table(x[[1]])
-#
-# # (?:^M|K|R|^)(WLSPEEVL)(?=[^P]|$)
-#
-# sequence[1:10]
-# fasta_vec[x[[1]]]
diff --git a/R/helpers-collapse_nodes_edgelist.R b/R/helpers-collapse_nodes_edgelist.R
@@ -1,4 +1,3 @@
-### aus Studienprojekt WS 22/23
 #' Collapsing of peptide and protein nodes of an edgelist.
 #'
 #' @param edgelist edgelist
@@ -18,9 +17,6 @@
 #'
 
 
-### TODO: Die Funktion funktioniert derzeit noch nicht, wenn z.B. die Proteinknoten schon collapsed sind!
-
-
 collapse_edgelist <- function(edgelist,
                               collapse_protein_nodes = TRUE,
                               collapse_peptide_nodes = TRUE) {
@@ -57,11 +53,11 @@ collapse_edgelist <- function(edgelist,
   #keep <- logical(nrow(edgelist2))
 
   pepNodes2 <- pepNodes
-  pepNodes2$peptide <- limma::strsplit2(pepNodes2$peptide, ";")[,1]  # erstes Peptid aus Liste!
+  pepNodes2$peptide <- limma::strsplit2(pepNodes2$peptide, ";")[,1]  # first peptide from list
   edgelist2 <- edgelist[edgelist$peptide %in% pepNodes2$peptide,]
 
   protNodes2 <- protNodes
-  protNodes2$protein <- limma::strsplit2(protNodes2$protein, ";")[,1]  # erstes Protein aus Liste!
+  protNodes2$protein <- limma::strsplit2(protNodes2$protein, ";")[,1]  # first peptide from list
   edgelist3 <- edgelist2[edgelist2$protein %in% protNodes2$protein,]
 
   edgelist4 <- edgelist3

diff --git a/R/helpers-collapse_nodes_edgelist_quant_pepratio.R b/R/helpers-collapse_nodes_edgelist_quant_pepratio.R
@@ -1,4 +1,3 @@
-### aus Studienprojekt WS 22/23
 #' Collapsing of peptide and protein nodes of an edgelist.
 #'
 #' @param edgelist edgelist
@@ -18,9 +17,6 @@
 #'
 
 
-### TODO: Die Funktion funktioniert derzeit noch nicht, wenn z.B. die Proteinknoten schon collapsed sind!
-
-
 collapse_edgelist_quant <- function(edgelist,
                               collapse_protein_nodes = TRUE,
                               collapse_peptide_nodes = TRUE) {
@@ -54,16 +50,14 @@ collapse_edgelist_quant <- function(edgelist,
 
 
   edgelist2 <- edgelist
-  #keep <- logical(nrow(edgelist2))
-
 
   pepNodes2 <- pepNodes
-  pepNodes2$peptide <- limma::strsplit2(pepNodes2$peptide, ";")[,1]  # erstes Peptid aus Liste!
+  pepNodes2$peptide <- limma::strsplit2(pepNodes2$peptide, ";")[,1]  # first peptide from list
   edgelist2 <- edgelist[edgelist$peptide %in% pepNodes2$peptide,]
 
 
   protNodes2 <- protNodes
-  protNodes2$protein <- limma::strsplit2(protNodes2$protein, ";")[,1]  # erstes Protein aus Liste!
+  protNodes2$protein <- limma::strsplit2(protNodes2$protein, ";")[,1]  # first peptide from list
   edgelist3 <- edgelist2[edgelist2$protein %in% protNodes2$protein,]
 
   edgelist4 <- edgelist3

diff --git a/R/helpers-convertToBipartiteGraph.R b/R/helpers-convertToBipartiteGraph.R
@@ -1,5 +1,4 @@
-
-#' Conversion of submatrizes to subgraphs.
+#' Conversion of submatrices to subgraphs.
 #'
 #' @param x element of a submatrix list
 #'

diff --git a/R/helpers-generate_edgelist.R b/R/helpers-generate_edgelist.R
@@ -1,6 +1,3 @@
-
-
-
 #' Generate edgelist from list of in silico digested proteins
 #'
 #' @param digested_proteins Output from digest_fasta() (List of vectors of peptide sequences)
@@ -16,9 +13,6 @@
 #' edgelist <- generate_edgelist(digested_proteins)
 #'
 #'
-#'
-#'
-#'
 generate_edgelist <- function(digested_proteins, prot_origin = NULL) {
   #calculate necessary number of edges by counting the peptides belonging to each protein
   mat_length <- sum(lengths(digested_proteins))

diff --git a/R/helpers-generate_graphs_via_edgelist.R b/R/helpers-generate_graphs_via_edgelist.R
@@ -1,5 +1,3 @@
-
-
 #' Generate bipartite peptide-protein graphs from a list of digested proteins via an edgelist
 #'
 #' @param edgelist Output from generate_edgelist (edgelist)
@@ -16,8 +14,6 @@
 #' res <- generate_graphs_from_edgelist(edgelist)
 #'
 
-## TODO: weitere Spalten in Edgelist (z.B. protein_origin)
-
 generate_graphs_from_edgelist <- function(edgelist) {
 
   #generate graph from edge matrix
@@ -26,7 +22,6 @@ generate_graphs_from_edgelist <- function(edgelist) {
   #assign vertex types to proteins and peptides for the graph to be bipartite
   igraph::V(G)[igraph::V(G)$name %in% edgelist[,1]]$type <- TRUE
   igraph::V(G)[igraph::V(G)$name %in% edgelist[,2]]$type <- FALSE
-  ### TODO: export G
 
   #decompose graph into connected components
   subgraphs <- igraph::decompose(G)

diff --git a/R/helpers-isomorphisms.R b/R/helpers-isomorphisms.R
@@ -8,64 +8,25 @@
 #' @return TRUE if graphs are isomorphic, FALSE if not.
 #' @export
 #'
-#' @examples # TODO
+#' @examples
 #'
 #'
 
-#graph1 <- G4
-#graph2 <- G5
-
-
 isomorphic_bipartite <- function(graph1, graph2, ...) {
 
-  #iso <- igraph::isomorphic(graph1, graph2)
-
   ## direct graphs if they are not directed yet
   if (!igraph::is_directed(graph1))   graph1 <- bppg::direct_bipartite_graph(graph1)
   if (!igraph::is_directed(graph2))   graph2 <- bppg::direct_bipartite_graph(graph2)
-  #graph2 <- bppg::direct_bipartite_graph(graph2)
-
 
- # igraph::V(graph1)$color <- c("black", "white")[igraph::V(graph1)$type+1]
-#  igraph::V(graph2)$color <- c("black", "white")[igraph::V(graph2)$type+1]
+  iso <- igraph::isomorphic(graph1, graph2, method = "vf2")
 
-
-  iso <- igraph::isomorphic(graph1, graph2, method = "vf2")#, vertex.color1 = igraph::V(graph1)$color, vertex.color2 = igraph::V(graph2)$color)
-
-
-  # if(iso) {
-  #   ## list all attributes except "type" to remove them before comparing the graphs
-  #
-  #
-  #   cG1 <- igraph::canonical_permutation(graph1)#, colors  = igraph::V(graph1)$type)
-  #   cG1 <- igraph::permute(graph1, cG1$labeling)
-  #
-  #   cG2 <- igraph::canonical_permutation(graph2)#, colors  = igraph::V(graph2)$type)
-  #   cG2 <- igraph::permute(graph2, cG2$labeling)
-  #
-  #
-  #   attribute_list <- unique(c(igraph::vertex_attr_names(graph1), igraph::vertex_attr_names(graph2)))
-  #   attribute_list <- attribute_list[!(attribute_list %in% c("type", "color"))]
-  #
-  #
-  #   # if there are any attributes other than "type", they will be removed
-  #   if (length(attribute_list) > 0) {
-  #     for (i in 1:length(attribute_list)) {
-  #       cG1 <- try({delete_vertex_attr(cG1, name = attribute_list[[i]])})
-  #       cG2 <- try({delete_vertex_attr(cG2, name = attribute_list[[i]])})
-  #     }
-  #   }
-  #
-  #   iso <- igraph::identical_graphs(cG1, cG2, attrs = FALSE)#all(igraph::V(cG1)$type == igraph::V(cG2)$type)
-  #
-  # }
   return(iso)
 }
 
 
 
 
-#' Title
+#' Transform a bipartite graph into a directed graph
 #'
 #' @param bip_graph
 #' @param from_type TODO
@@ -74,7 +35,7 @@ isomorphic_bipartite <- function(graph1, graph2, ...) {
 #' @export
 #'
 #' @examples
-#' # TODO
+#'
 direct_bipartite_graph <- function(bip_graph, from_type = FALSE){
 
 

diff --git a/R/helpers-preprocess_quant_peptide_data.R b/R/helpers-preprocess_quant_peptide_data.R
@@ -1,9 +1,3 @@
-
-
-### TODO: read in data directly from MaxQuant and filter unnecessary columns and decoys
-### TODO: Normalization
-
-
 #' Import of MaxQuant's peptide.txt-table
 #'
 #' @param path Path to the peptides.txt table
@@ -112,10 +106,8 @@ aggregate_replicates <- function(D, group, missing.limit = 0, method = "mean",
 
     res_tmp <- FUN(X_tmp, na.rm = TRUE)
 
-   # if (!use0) {
-      missingx <- apply(X_tmp, 1, function(x) mean(is.na(x)))
-      res_tmp[missingx > missing.limit | missingx == 1] <- NA
-  #  }
+    missingx <- apply(X_tmp, 1, function(x) mean(is.na(x)))
+    res_tmp[missingx > missing.limit | missingx == 1] <- NA
 
     res <- cbind(res, res_tmp)
   }
@@ -139,7 +131,7 @@ aggregate_replicates <- function(D, group, missing.limit = 0, method = "mean",
 #' @export
 #'
 #' @examples
-#' ### TODO
+#'
 foldChange <- function(D, X, Y, useNA = FALSE) {
   FC <- D[, Y] / D[, X]
 
@@ -164,7 +156,7 @@ foldChange <- function(D, X, Y, useNA = FALSE) {
 #' @export
 #'
 #' @examples
-#' ## TODO
+#'
 calculate_peptide_ratios <- function(aggr_intensities, id_cols = 1,
                                      group_levels = NULL, type = "ratio", log_base = 10) {