Build Adjacency Matrix

#R

file_path <- file.choose()
data_matrix <- read.table(file_path, header = TRUE)

A <- 1:4756  # Define the range of column numbers to process

for (col_num in A) {
  p_values <- data_matrix[-1, col_num]
  
  sorted_p_values <- sort(p_values)
  
  top_500_p_values <- head(sorted_p_values, 500)
  
  indices <- which(data_matrix[-1, col_num] %in% top_500_p_values)
  
  related_names <- data_matrix[indices + 1, 1]  # Adding 1 to indices due to header row
  
  X <- gconvert(query = related_names, organism = "hsapiens", target = "ENSG")
  
  proteins_mapped <- rba_string_map_ids(ids = X$name, species = 9606)
  
  all_proteins <- unique(proteins_mapped[, 2])
  
  int_net <- rba_string_interactions_network(
    ids = proteins_mapped$stringId,
    species = 9606,
    network_type = "functional",
    required_score = 900
  )
  
  matrix <- int_net[, c(3, 4, 6)]
  
  adj_matrix <- matrix(0, nrow = length(all_proteins), ncol = length(all_proteins))
  rownames(adj_matrix) <- all_proteins
  colnames(adj_matrix) <- all_proteins
  
  for (i in 1:nrow(matrix)) {
    protein1 <- matrix[i, 1]
    protein2 <- matrix[i, 2]
    
    if (protein1 %in% all_proteins && protein2 %in% all_proteins) {
      adj_matrix[protein1, protein2] <- matrix[i, 3]
      adj_matrix[protein2, protein1] <- matrix[i, 3]  # Assuming an undirected graph
    }
  }
  
  file_name <- paste0("adj_matrix_col_num_", col_num, ".csv")
  
  write.csv(adj_matrix, file = file_name, row.names = TRUE)
  
  print(paste("Adjacency matrix saved to", file_name))
}