renamed files with shorter names and corrected paths for input and output files

mcsm_analysis/pyrazinamide/scripts/mcsm/run.sh

@@ -1,6 +1,9 @@
-# run step0-step3a for mcsm pipeline
+#!/bin/bash
+
+# run all bash scripts for mcsm
 
 #./step0_check_duplicate_SNPs.sh
-#./step1_mCSMLig_curl_submit_store_outputurl.sh
+#./step1_lig_output_urls.sh
-./step2_mCSM_LIG_batch_outputurls_results.sh
+./step2_lig_results.sh
-./step3a_mCSM_LIG_regex_output_formatting.sh
+./step3a_results_format_interim.sh
+

mcsm_analysis/pyrazinamide/scripts/mcsm/step1_lig_output_urls.sh

@@ -24,7 +24,7 @@ infile_mut="/pnca_mis_SNPs_v2_unique.csv"
 infile_struc="/complex1_no_water.pdb"
 
 outpath="${inpath}${processed_path}"
-outfile="/mCSM_lig_complex1_result_url.txt"
+outfile="/complex1_result_url.txt"
 
 # create valid input and output filenames
 #filename="${HOME}/git/Data/pyrazinamide/input/processed/pnca_mis_SNPs_v2_unique.csv"
@@ -82,7 +82,7 @@ echo -e "${mutation} : processing entry ${COUNT}/$(wc -l < ${filename})..."
 
 # create output file with the added number of muts from file
 # after much thought, bad idea as less generic!
-#echo -e "${host}${result_url}" >> ../Results/$(wc -l < ${filename})_mCSM_lig_complex1_result_url.txt
+#echo -e "${host}${result_url}" >> ../Results/$(wc -l < ${filename})_complex1_result_url.txt
 echo -e "${host}${result_url}" >> ${outfilename}
 #echo -n '.'
 done < "${filename}"

mcsm_analysis/pyrazinamide/scripts/mcsm/step2_lig_results.sh

@@ -35,10 +35,10 @@
 # specify variables for input and output paths and filenames
 inpath="${HOME}/git/Data/pyrazinamide/input"
 processed_path="/processed"
-infile="/mCSM_lig_complex1_result_url.txt"
+infile="/complex1_result_url.txt"
 
 outpath="${inpath}${processed_path}"
-outfile="/mCSM_lig_complex1_output_MASTER.txt"
+outfile="/complex1_output_MASTER.txt"
 
 # create valid input and output filenames
 filename="${inpath}${processed_path}${infile}"

mcsm_analysis/pyrazinamide/scripts/mcsm/step3a_results_format_interim.sh

@@ -27,12 +27,11 @@ inpath="${HOME}/git/Data/pyrazinamide/input"
 processed_path="/processed"
 
 # Create input file: copy and rename output file of step2
-oldfile="${inpath}${processed_path}/mCSM_lig_complex1_output_MASTER.txt"
+oldfile="${inpath}${processed_path}/complex1_output_MASTER.txt"
-newfile="${inpath}${processed_path}/mCSM_lig_complex1_output_processed.txt"
+newfile="${inpath}${processed_path}/complex1_output_processed.txt"
 cp $oldfile $newfile
 
-#infile="../Results/336_mCSM_lig_complex1_output_processed.txt"
+infile="/complex1_output_processed.txt"
-infile="/mCSM_lig_complex1_output_processed.txt"
 filename="${inpath}${processed_path}${infile}"
 
 echo Input filename is : ${filename}

mcsm_analysis/pyrazinamide/scripts/mcsm/step3b_format_results.py

@@ -1,29 +0,0 @@
-#!/usr/bin/python
-import pandas as pd
-from collections import defaultdict
-
-#file = r'../Results/322_mCSM_lig_complex1_output_processed.txt'
-
-outCols=[
-        'PredictedAffinityChange',
-        'Mutationinformation',
-        'Wild-type',
-        'Position',
-        'Mutant-type',
-        'Chain',
-        'LigandID',
-        'Distancetoligand',
-        'DUETstabilitychange'
-        ]
-
-lines = [line.rstrip('\n') for line in open('../Results/336_mCSM_lig_complex1_output_processed.txt')]
-
-outputs = defaultdict(list)
-
-for item in lines:
-	col, val = item.split(':')
-	outputs[col].append(val)
-
-dfOut=pd.DataFrame(outputs)
-
-pd.DataFrame.to_csv(dfOut,'../Results/336_complex1_formatted_results.csv', columns=outCols)

mcsm_analysis/pyrazinamide/scripts/mcsm/step3b_results_format_df.py

@@ -0,0 +1,61 @@
+#!/usr/bin/python
+
+###################
+# load libraries
+import os, sys
+import pandas as pd
+from collections import defaultdict
+####################
+
+#********************************************************************
+# TASK: Formatting results with nice colnames
+# step3a processed the mcsm results to remove all newlines and 
+# brought data in a format where the delimiter ":" splits
+# data into a convenient format of "colname": "value".
+# this script formats the data and outputs a df with each row
+# as a mutation and its corresponding mcsm_values
+
+# Requirements:
+# input: output of step3a, file containing  "..._output_processed.txt"
+	# path: "Data/<drug>/input/processed/<filename>"
+# output: formatted .csv file
+	# path: "Data/<drug>/input/processed/<filename>"
+#***********************************************************************
+# specify variables for input and output paths and filenames
+homedir = os.path.expanduser('~') # spyder/python doesn't recognise tilde
+
+basedir = "/git/Data/pyrazinamide/input"
+inpath = "/processed"
+in_filename = "/complex1_output_processed.txt"
+infile = homedir + basedir + inpath + in_filename
+print("Input file is:", infile)
+
+outpath = "/processed"
+out_filename = "/complex1_formatted_results.csv"
+outfile = homedir + basedir + outpath + out_filename
+print("Output file is:", outfile)
+# end of variable assignment for input and output files
+
+outCols=[
+        'PredictedAffinityChange',
+        'Mutationinformation',
+        'Wild-type',
+        'Position',
+        'Mutant-type',
+        'Chain',
+        'LigandID',
+        'Distancetoligand',
+        'DUETstabilitychange'
+        ]
+
+lines = [line.rstrip('\n') for line in open(infile)]
+
+outputs = defaultdict(list)
+
+for item in lines:
+	col, val = item.split(':')
+	outputs[col].append(val)
+
+dfOut=pd.DataFrame(outputs)
+
+pd.DataFrame.to_csv(dfOut, outfile, columns=outCols)

mcsm_analysis/pyrazinamide/scripts/mcsm/step3c_results_cleaning.R

@@ -1,22 +1,42 @@
 getwd()
-#setwd("~/Documents/git/LSHTM_analysis/mcsm_complex1/Results") # work
+#setwd("~/git/LSHTM_analysis/mcsm_complex1/Results")
-setwd("~/git/LSHTM_analysis/mcsm_complex1/Results") # thinkpad
-#setwd("/Users/tanu/git/LSHTM_analysis/mcsm_complex1/Results") # mac
 getwd()
 
 #=======================================================
+# TASK: read formatted_results_df.csv to complete 
+# missing info, adding DUET categories, assigning
+# meaningful colnames, etc.
+
+# Requirements:
+# input: output of step3b, python processing,
+  # path: Data/<drug>/input/processed/<filename>"
+# output: NO output as the next scripts refers to this
+# for yet more processing
+#=======================================================
+
+# specify variables for input and output paths and filenames
+homedir = "~"
+basedir = "/git/Data/pyrazinamide/input"
+inpath = "/processed"
+in_filename = "/complex1_formatted_results.csv"
+infile = paste0(homedir, basedir, inpath, in_filename)
+print(paste0("Input file is:", infile))
+
+#======================================================
 #TASK: To tidy the columns so you can generate figures
 #=======================================================
 ####################
 #### read file #####: this will be the output from python script (csv file)
 ####################
-data = read.csv("336_complex1_formatted_results.csv"
+data = read.csv(infile
               , header = T
               , stringsAsFactors = FALSE)
 dim(data)
-#335, 10
 str(data)
 
+# clear variables
+rm(homedir, basedir, inpath, in_filename, infile)
+
 ###########################
 ##### Data processing #####
 ###########################
@@ -31,34 +51,30 @@ data$Mutationinformation = paste0(data$Wild.type, data$Position, data$Mutant.typ
 head(data$Mutationinformation)
 tail(data$Mutationinformation)
 #write.csv(data, 'test.csv')
+
 ##########################################
 # Remove duplicate SNPs as a sanity check
 ##########################################
 # very important
 table(duplicated(data$Mutationinformation))
-#FALSE   
-#335
 
 # extract duplicated entries
 dups = data[duplicated(data$Mutationinformation),] #0
 
 # No of dups should match with the no. of TRUE in the above table 
 #u_dups = unique(dups$Mutationinformation) #10
-sum( table(dups$Mutationinformation) ) #13
+sum( table(dups$Mutationinformation) )
-
-rm(dups)
 
 #***************************************************************
 # select non-duplicated SNPs and create a new df
-df = data[!duplicated(data$Mutationinformation),] #309, 10
+df = data[!duplicated(data$Mutationinformation),]
 #***************************************************************
 # sanity check
 u = unique(df$Mutationinformation)
 u2 = unique(data$Mutationinformation)
 table(u%in%u2)
-#TRUE 
+
-#309 
+# should all be 1
-#should all be 1, hence 309 1's
 sum(table(df$Mutationinformation) == 1)
 
 # sort df by Position
@@ -66,6 +82,7 @@ sum(table(df$Mutationinformation) == 1)
 #foo <- df[order(df$Position),]
 #df <- df[order(df$Position),]
 
+# clear variables
 rm(u, u2, dups)
 
 ####################
@@ -76,7 +93,7 @@ rm(u, u2, dups)
 #STEP 1
 #========
 # make a copy of the PredictedAffinityColumn and call it Lig_outcome
-df$Lig_outcome = df$PredictedAffinityChange #335, 11
+df$Lig_outcome = df$PredictedAffinityChange
 
  #make Predicted...column numeric and outcome column categorical
 head(df$PredictedAffinityChange)
@@ -89,8 +106,10 @@ head(df$PredictedAffinityChange)
 
 # should be numeric, check and if not make it numeric
 is.numeric( df$PredictedAffinityChange )
+
 # change to numeric
 df$PredictedAffinityChange = as.numeric(df$PredictedAffinityChange)
+
 # should be TRUE
 is.numeric( df$PredictedAffinityChange )
 
@@ -109,10 +128,13 @@ df$Lig_outcome = gsub("^.*-"
                   df$Lig_outcome)
 # sanity checks
 head(df$Lig_outcome)
+
 # should be factor, check and if not change it to factor
 is.factor(df$Lig_outcome) 
+
 # change to factor
 df$Lig_outcome = as.factor(df$Lig_outcome)
+
 # should be TRUE
 is.factor(df$Lig_outcome) 
 
@@ -126,10 +148,13 @@ df$Distancetoligand = gsub("&Aring;"
                            , df$Distancetoligand)
 # sanity checks
 head(df$Distancetoligand)
+
 # should be numeric, check if not change it to numeric
 is.numeric(df$Distancetoligand)
+
 # change to numeric
 df$Distancetoligand = as.numeric(df$Distancetoligand)
+
 # should be TRUE
 is.numeric(df$Distancetoligand)
 
@@ -148,10 +173,13 @@ df$DUETstabilitychange = gsub("Kcal/mol"
                               , df$DUETstabilitychange)
 # sanity checks
 head(df$DUETstabilitychange)
+
 # should be numeric, check if not change it to numeric
 is.numeric(df$DUETstabilitychange)
+
 # change to numeric 
 df$DUETstabilitychange = as.numeric(df$DUETstabilitychange)
+
 # should be TRUE
 is.numeric(df$DUETstabilitychange)
 
@@ -166,22 +194,17 @@ colnames(df)[n]
 # create yet another extra column: classification of DUET stability only
 df$DUET_outcome = ifelse(df$DUETStability_Kcalpermol >=0
                          , "Stabilizing"
-                         , "Destabilizing")  #335, 12
+                         , "Destabilizing") # spelling to be consistent with mcsm
 
 table(df$Lig_outcome)
-#Destabilizing   Stabilizing 
-#281             54 
 
 table(df$DUET_outcome)
-#Destabilizing   Stabilizing 
+
-#288             47 
 #==============================
 #FIXME
 #Insert a venn diagram
-
 #================================
 
-
 #========
 #STEP 6
 #========
@@ -199,7 +222,7 @@ colnames(df[mut])
 #STEP 7
 #========
 # create an extra column: maybe useful for some plots
-df$WildPos = paste0(df$Wild_type, df$Position) #335, 13
+df$WildPos = paste0(df$Wild_type, df$Position)
 
 # clear variables
 rm(n, wt, mut)

mcsm_analysis/pyrazinamide/scripts/mcsm/step4_results_normalise.R

@@ -1,7 +1,8 @@
-getwd()
+##################
-#setwd("~/Documents/git/LSHTM_analysis/mcsm_complex1/Results") # work
+# load libraries
-setwd("~/git/LSHTM_analysis/mcsm_complex1/Results") # thinkpad
+ library(compare)
-#setwd("/Users/tanu/git/LSHTM_analysis/mcsm_complex1/Results") # mac
+##################
+
 getwd()
 
 #=======================================================
@@ -9,17 +10,38 @@ getwd()
   # of DUET stability scores
   # of Pred affinity
 # compare scaling methods with plots
+
+# Requirements:
+# input: R script, step3c_results_cleaning.R
+  # path: Data/<drug>/input/processed/<filename>"
+# output: NO output as the next scripts refers to this
+# for yet more processing
 # output normalised file
 #=======================================================
 
+# specify variables for input and output paths and filenames
+homedir = "~"
+currdir = "/git/LSHTM_analysis/mcsm_analysis/pyrazinamide/scripts/mcsm"
+in_filename = "/step3c_results_cleaning.R"
+infile = paste0(homedir, currdir, in_filename)
+print(paste0("Input file is:", infile))
+
+# output file
+basedir = "/git/Data/pyrazinamide/input"
+outpath = "/processed"
+out_filename = "/mcsm_complex1_normalised.csv"
+outfile = paste0(homedir, basedir, outpath, out_filename)
+print(paste0("Output file is:", outfile))
+
 ####################
 #### read file #####: this will be the output of my R script that cleans the data columns
 ####################
-source("../Scripts/step3c_data_cleaning.R")
+source(infile)
-##This will outut two dataframes:
+
-##data: unclean data: 335, 10
+#This will outut two dataframes:
-##df : cleaned df 335, 13
+# data: unclean data: 10 cols
-## you can remove data if you want as you will not need it
+# df : cleaned df: 13 cols
+# you can remove data if you want as you will not need it
 rm(data)
 
 colnames(df)
@@ -49,26 +71,25 @@ mean(df[,n])
 #-0.9526746
 
 tapply(df[,n], df[,group], mean)
-#Destabilizing   Stabilizing 
+
-#-1.2112100      0.3926667 
 #===========================
 # Same as above: in 2 steps
 #===========================
 
 # find range of your data
-my_min = min(df[,n]); my_min #-3.948
+my_min = min(df[,n]); my_min #
-my_max = max(df[,n]); my_max #2.23
+my_max = max(df[,n]); my_max #
 
 #===============================================
 # WITHIN GROUP rescaling 2: method "ratio"
 # create column to store the rescaled values
 # Rescaling separately (Less dangerous) 
-#       =====> chosen one:as Nick prefers
+#       =====> chosen one: preserves sign
 #===============================================
 df$ratioPredAff = ifelse(df[,n] < 0
                       , df[,n]/abs(my_min)
                       , df[,n]/my_max
-                      )#335 14
+                      )# 14 cols
 # sanity checks
 head(df$ratioPredAff)
 tail(df$ratioPredAff)
@@ -76,19 +97,13 @@ tail(df$ratioPredAff)
 min(df$ratioPredAff); max(df$ratioPredAff)
 
 tapply(df$ratioPredAff, df$Lig_outcome, min)
-#Destabilizing   Stabilizing 
-#-1.000000000   0.005381166 
 
 tapply(df$ratioPredAff, df$Lig_outcome, max)
-#Destabilizing   Stabilizing 
-#-0.001266464   1.000000000
 
-#should be the same as below (281 and 54)
+# should be the same as below 
 sum(df$ratioPredAff < 0); sum(df$ratioPredAff > 0)
 
 table(df$Lig_outcome)
-#Destabilizing   Stabilizing 
-#281              54
 
 #===============================================
 # Hist and density plots to compare the rescaling 
@@ -140,7 +155,7 @@ rm(my_min, my_max, my_title, n, group)
 #===================
 # 3b: DUET stability
 #===================
-dim(df) #335, 14
+dim(df) # 14 cols
 
 n = which(colnames(df) == "DUETStability_Kcalpermol"); n # 10
 group = which(colnames(df) == "DUET_outcome"); group #12
@@ -160,26 +175,23 @@ tail(df[,n]) #positives
 
 # sanity checks
 mean(df[,n])
-#[1] -1.173316
 
 tapply(df[,n], df[,group], mean)
-#Destabilizing   Stabilizing 
-#-1.4297257     0.3978723
 
 #===============================================
 # WITHIN GROUP rescaling 2: method "ratio"
 # create column to store the rescaled values
 # Rescaling separately (Less dangerous) 
-#       =====> chosen one:as Nick prefers
+#       =====> chosen one: preserves sign
 #===============================================
 # find range of your data
-my_min = min(df[,n]); my_min #-3.87
+my_min = min(df[,n]); my_min 
-my_max = max(df[,n]); my_max #1.689
+my_max = max(df[,n]); my_max
 
 df$ratioDUET = ifelse(df[,n] < 0
                       , df[,n]/abs(my_min)
                       , df[,n]/my_max
-                    ) #335, 15
+                    ) # 15 cols
 # sanity check
 head(df$ratioDUET)
 tail(df$ratioDUET)
@@ -188,26 +200,19 @@ min(df$ratioDUET); max(df$ratioDUET)
 
 # sanity checks
 tapply(df$ratioDUET, df$DUET_outcome, min)
-#Destabilizing   Stabilizing 
-#-1.00000000    0.01065719
 
 tapply(df$ratioDUET, df$DUET_outcome, max)
-#Destabilizing   Stabilizing 
-#-0.003875969   1.000000000 
 
 # should be the same as below (267 and 42)
 sum(df$ratioDUET < 0); sum(df$ratioDUET > 0)
 
 table(df$DUET_outcome)
-#Destabilizing   Stabilizing 
-#288             47
 
 #===============================================
 # Hist and density plots to compare the rescaling 
 # methods: Base R
 #===============================================
 # uncomment as necessary
-
 my_title = "DUET_stability"
 #my_title = colnames(df[n])
 
@@ -246,7 +251,25 @@ mtext(text = my_title
       , line = 0
       , outer = TRUE)
 
+# reorder by column name
+#data <- data[c("A", "B", "C")]
+colnames(df)
+df2 = df[c("X", "Mutationinformation",  "WildPos", "Position"
+           , "Wild_type", "Mutant_type"
+           , "DUETStability_Kcalpermol", "DUET_outcome"
+           , "Dis_lig_Ang", "PredAffLog", "Lig_outcome"
+           , "ratioDUET", "ratioPredAff"
+           , "LigandID","Chain")]
+
+# sanity check
+# should be True
+#compare(df, df2, allowAll = T)
+compare(df, df2, ignoreColOrder = T)
+#TRUE 
+#reordered columns
+
 #===================
 # write output as csv file
 #===================
-write.csv(df, "../Data/mcsm_complex1_normalised.csv", row.names = FALSE) #335, 15
+#write.csv(df, "../Data/mcsm_complex1_normalised.csv", row.names = FALSE)
+write.csv(df2, outfile, row.names = FALSE)