LSHTM_analysis/meta_data_analysis/combining_df_ps.R

461 lines
16 KiB
R

#=======================================================================
# TASK: To combine mcsm and meta data with af and or files
# Input csv files:
# 1) mcsm normalised and struct params
# 2) gene associated meta_data_with_AFandOR
# Output:
# 1) muts with opposite effects on stability
# 2) large combined df including NAs for AF, OR,etc
# Dim: same no. of rows as gene associated meta_data_with_AFandOR
# 3) small combined df including NAs for AF, OR, etc.
# Dim: same as mcsm data
# 4) large combined df excluding NAs
# Dim: dim(#1) - no. of NAs(AF|OR) + 1
# 5) small combined df excluding NAs
# Dim: dim(#2) - no. of unique NAs - 1
# This script is sourced from other .R scripts for plotting
#=======================================================================
#%% specify curr dir
getwd()
setwd('~/git/LSHTM_analysis/meta_data_analysis/')
getwd()
#=======================================================================
#%% load packages
#require(data.table)
#require(arsenal)
#require(compare)
#library(tidyverse)
# header file
header_dir = '~/git/LSHTM_analysis/'
source(paste0(header_dir, '/', 'my_header.R'))
#=======================================================================
#%% variable assignment: input and output paths & filenames
drug = 'pyrazinamide'
gene = 'pncA'
gene_match = paste0(gene,'_p.')
cat(gene_match)
#===========
# data dir
#===========
datadir = '~/git/Data'
#===========
# input
#===========
# infile1: mCSM data
#indir = '~/git/Data/pyrazinamide/input/processed/'
indir = paste0(datadir, '/', drug, '/', 'output') # revised {TODO: change in mcsm pipeline}
#in_filename = 'mcsm_complex1_normalised.csv'
in_filename = 'pnca_mcsm_struct_params.csv'
infile = paste0(indir, '/', in_filename)
cat(paste0('Reading infile1: mCSM output file', ' ', infile, '\n') )
# infile2: gene associated meta data combined with AF and OR
#indir: same as above
in_filename_comb = paste0(tolower(gene), '_meta_data_with_AFandOR.csv')
infile_comb = paste0(indir, '/', in_filename_comb)
cat(paste0('Reading infile2: gene associated combined metadata:', infile_comb, '\n'))
#===========
# output
#===========
# Uncomment if and when required to output
outdir = paste0(datadir, '/', drug, '/', 'output') #same as indir
cat('Output dir: ', outdir, '\n')
#out_filename = paste0(tolower(gene), 'XXX')
#outfile = paste0(outdir, '/', out_filename)
#cat(paste0('Output file with full path:', outfile))
#%% end of variable assignment for input and output files
#=======================================================================
#%% Read input files
#####################################
# input file 1: mcsm normalised data
# output of step 4 mcsm_pipeline
#####################################
cat('Reading mcsm_data:'
, '\nindir: ', indir
, '\ninfile_comb: ', in_filename)
mcsm_data = read.csv(infile
# , row.names = 1
, stringsAsFactors = F
, header = T)
cat('Read mcsm_data file:'
, '\nNo.of rows: ', nrow(mcsm_data)
, '\nNo. of cols:', ncol(mcsm_data))
# clear variables
rm(in_filename, infile)
str(mcsm_data)
table(mcsm_data$DUET_outcome); sum(table(mcsm_data$DUET_outcome) )
# spelling Correction 1: DUET
mcsm_data$DUET_outcome[mcsm_data$DUET_outcome=='Stabilizing'] <- 'Stabilising'
mcsm_data$DUET_outcome[mcsm_data$DUET_outcome=='Destabilizing'] <- 'Destabilising'
# checks: should be the same as above
table(mcsm_data$DUET_outcome); sum(table(mcsm_data$DUET_outcome) )
head(mcsm_data$DUET_outcome); tail(mcsm_data$DUET_outcome)
# spelling Correction 2: Ligand
table(mcsm_data$Lig_outcome); sum(table(mcsm_data$Lig_outcome) )
mcsm_data$Lig_outcome[mcsm_data$Lig_outcome=='Stabilizing'] <- 'Stabilising'
mcsm_data$Lig_outcome[mcsm_data$Lig_outcome=='Destabilizing'] <- 'Destabilising'
# checks: should be the same as above
table(mcsm_data$Lig_outcome); sum(table(mcsm_data$Lig_outcome) )
head(mcsm_data$Lig_outcome); tail(mcsm_data$Lig_outcome)
# muts with opposing effects on protomer and ligand stability
table(mcsm_data$DUET_outcome != mcsm_data$Lig_outcome)
changes = mcsm_data[which(mcsm_data$DUET_outcome != mcsm_data$Lig_outcome),]
# sanity check: redundant, but uber cautious!
dl_i = which(mcsm_data$DUET_outcome != mcsm_data$Lig_outcome)
ld_i = which(mcsm_data$Lig_outcome != mcsm_data$DUET_outcome)
cat('Identifying muts with opposite stability effects')
if(nrow(changes) == (table(mcsm_data$DUET_outcome != mcsm_data$Lig_outcome)[[2]]) & identical(dl_i,ld_i)) {
cat('PASS: muts with opposite effects on stability and affinity correctly identified'
, '\nNo. of such muts: ', nrow(changes))
}else {
cat('FAIL: unsuccessful in extracting muts with changed stability effects')
}
#***************************
# write file: changed muts
out_filename = 'muts_opp_effects.csv'
outfile = paste0(outdir, '/', out_filename)
cat('Writing file for muts with opp effects:'
, '\nFilename: ', outfile
, '\nPath: ', outdir)
write.csv(changes, outfile)
#****************************
# clear variables
rm(out_filename, outfile)
rm(changes, dl_i, ld_i)
# count na in each column
na_count = sapply(mcsm_data, function(y) sum(length(which(is.na(y))))); na_count
# sort by Mutationinformation
mcsm_data = mcsm_data[order(mcsm_data$Mutationinformation),]
head(mcsm_data$Mutationinformation)
orig_col = ncol(mcsm_data)
# get freq count of positions and add to the df
setDT(mcsm_data)[, mut_pos_occurrence := .N, by = .(Position)]
cat('Added col: position frequency of muts to see which posn has how many muts'
, '\nNo. of cols now', ncol(mcsm_data)
, '\nNo. of cols before: ', orig_col)
mut_pos_occurrence = data.frame(mcsm_data$Mutationinformation
, mcsm_data$Position
, mcsm_data$mut_pos_occurrence)
colnames(mut_pos_occurrence) = c('Mutationinformation', 'position', 'mut_pos_occurrence')
#######################################
# input file 2: meta data with AFandOR
#######################################
cat('Reading combined meta data and AFandOR file:'
, '\nindir: ', indir
, '\ninfile_comb: ', in_filename_comb)
meta_with_afor <- read.csv(infile_comb
, stringsAsFactors = F
, header = T)
cat('Read mcsm_data file:'
, '\nNo.of rows: ', nrow(meta_with_afor)
, '\nNo. of cols:', ncol(meta_with_afor))
# counting NAs in AF, OR cols
if (identical(sum(is.na(meta_with_afor$OR))
, sum(is.na(meta_with_afor$pvalue))
, sum(is.na(meta_with_afor$AF)))){
cat('PASS: NA count match for OR, pvalue and AF\n')
na_count = sum(is.na(meta_with_afor$AF))
cat('No. of NAs: ', sum(is.na(meta_with_afor$OR)))
} else{
cat('FAIL: NA count mismatch'
, '\nNA in OR: ', sum(is.na(meta_with_afor$OR))
, '\nNA in pvalue: ', sum(is.na(meta_with_afor$pvalue))
, '\nNA in AF:', sum(is.na(meta_with_afor$AF)))
}
# clear variables
rm(in_filename_comb, infile_comb)
str(meta_with_afor)
# sort by Mutationinformation
head(meta_with_afor$Mutationinformation)
meta_with_afor = meta_with_afor[order(meta_with_afor$Mutationinformation),]
head(meta_with_afor$Mutationinformation)
orig_col2 = ncol(meta_with_afor)
# get freq count of positions and add to the df
setDT(meta_with_afor)[, sample_pos_occurrence := .N, by = .(position)]
cat('Added col: position frequency of samples to check'
,'how many samples correspond to a partiulcar posn associated with muts'
, '\nNo. of cols now', ncol(meta_with_afor)
, '\nNo. of cols before: ', orig_col2)
sample_pos_occurrence = data.frame(meta_with_afor$id
, meta_with_afor$mutation
, meta_with_afor$Mutationinformation
, meta_with_afor$position
, meta_with_afor$sample_pos_occurrence)
colnames(sample_pos_occurrence) = c('id', 'mutation', 'Mutationinformation', 'position', 'sample_pos_occurrence')
#=======================================================================
cat('Begin merging dfs with NAs'
, '\n===============================================================')
###########################
# merging two dfs: with NA
###########################
# link col name = 'Mutationinforamtion'
head(mcsm_data$Mutationinformation)
head(meta_with_afor$Mutationinformation)
cat('Merging dfs with NAs: big df (1-many relationship b/w id & mut)'
,'\nlinking col: Mutationinforamtion'
,'\nfilename: merged_df2')
#########
# a) merged_df2: meta data with mcsm
#########
merged_df2 = merge(x = meta_with_afor
,y = mcsm_data
, by = 'Mutationinformation'
, all.y = T)
cat('Dim of merged_df2: '
, '\nNo. of rows: ', nrow(merged_df2)
, '\nNo. of cols: ', ncol(merged_df2))
head(merged_df2$Position)
# sanity check
cat('Checking nrows in merged_df2')
if(nrow(meta_with_afor) == nrow(merged_df2)){
cat('nrow(merged_df2) = nrow (gene associated metadata)'
,'\nExpected no. of rows: ',nrow(meta_with_afor)
,'\nGot no. of rows: ', nrow(merged_df2))
} else{
cat('nrow(merged_df2)!= nrow(gene associated metadata)'
, '\nExpected no. of rows after merge: ', nrow(meta_with_afor)
, '\nGot no. of rows: ', nrow(merged_df2)
, '\nFinding discrepancy')
merged_muts_u = unique(merged_df2$Mutationinformation)
meta_muts_u = unique(meta_with_afor$Mutationinformation)
# find the index where it differs
unique(meta_muts_u[! meta_muts_u %in% merged_muts_u])
}
# sort by Position
head(merged_df2$Position)
merged_df2 = merged_df2[order(merged_df2$Position),]
head(merged_df2$Position)
merged_df2v2 = merge(x = meta_with_afor
,y = mcsm_data
, by = 'Mutationinformation'
, all.x = T)
#!=!=!=!=!=!=!=!
# COMMENT: used all.y since position 186 is not part of the struc,
# hence doesn't have a mcsm value
# but 186 is associated with mutation
#!=!=!=!=!=!=!=!
# should be False
identical(merged_df2, merged_df2v2)
table(merged_df2$Position%in%merged_df2v2$Position)
rm(merged_df2v2)
#########
# b) merged_df3:remove duplicate mutation information
#########
cat('Merging dfs without NAs: small df (removing muts with no AF|OR associated)'
,'\nCannot trust lineage info from this'
,'\nlinking col: Mutationinforamtion'
,'\nfilename: merged_df3')
#==#=#=#=#=#=#
# Cannot trust lineage, country from this df as the same mutation
# can have many different lineages
# but this should be good for the numerical corr plots
#=#=#=#=#=#=#=
merged_df3 = merged_df2[!duplicated(merged_df2$Mutationinformation),]
head(merged_df3$Position); tail(merged_df3$Position) # should be sorted
# sanity check
cat('Checking nrows in merged_df3')
if(nrow(mcsm_data) == nrow(merged_df3)){
cat('PASS: No. of rows match with mcsm_data'
,'\nExpected no. of rows: ', nrow(mcsm_data)
,'\nGot no. of rows: ', nrow(merged_df3))
} else {
cat('FAIL: No. of rows mismatch'
, '\nNo. of rows mcsm_data: ', nrow(mcsm_data)
, '\nNo. of rows merged_df3: ', nrow(merged_df3))
}
# counting NAs in AF, OR cols in merged_df3
# this is becuase mcsm has no AF, OR cols,
# so you cannot count NAs
if (identical(sum(is.na(merged_df3$OR))
, sum(is.na(merged_df3$pvalue))
, sum(is.na(merged_df3$AF)))){
cat('PASS: NA count match for OR, pvalue and AF\n')
na_count_df3 = sum(is.na(merged_df3$AF))
cat('No. of NAs: ', sum(is.na(merged_df3$OR)))
} else{
cat('FAIL: NA count mismatch'
, '\nNA in OR: ', sum(is.na(merged_df3$OR))
, '\nNA in pvalue: ', sum(is.na(merged_df3$pvalue))
, '\nNA in AF:', sum(is.na(merged_df3$AF)))
}
#=======================================================================
#%% merging without NAs
cat('Begin merging dfs without NAs'
, '\n===============================================================')
cat('Merging dfs without any NAs: big df (1-many relationship b/w id & mut)'
,'\nlinking col: Mutationinforamtion'
,'\nfilename: merged_df2_comp')
#########
# c) merged_df2_comp: same as merge 1 but excluding NA
#########
merged_df2_comp = merged_df2[!is.na(merged_df2$AF),]
#merged_df2_comp = merged_df2[!duplicated(merged_df2$Mutationinformation),]
# sanity check
cat('Checking nrows in merged_df2_comp')
if(nrow(merged_df2_comp) == (nrow(merged_df2) - na_count + 1)){
cat('PASS: No. of rows match'
,'\nDim of merged_df2_comp: '
,'\nExpected no. of rows: ', nrow(merged_df2) - na_count + 1
, '\nNo. of rows: ', nrow(merged_df2_comp)
, '\nNo. of cols: ', ncol(merged_df2_comp))
}else{
cat('FAIL: No. of rows mismatch'
,'\nExpected no. of rows: ', nrow(merged_df2) - na_count + 1
,'\nGot no. of rows: ', nrow(merged_df2_comp))
}
#########
# d) merged_df3_comp: remove duplicate mutation information
#########
merged_df3_comp = merged_df2_comp[!duplicated(merged_df2_comp$Mutationinformation),]
cat('Dim of merged_df3_comp: '
, '\nNo. of rows: ', nrow(merged_df3_comp)
, '\nNo. of cols: ', ncol(merged_df3_comp))
# alternate way of deriving merged_df3_comp
foo = merged_df3[!is.na(merged_df3$AF),]
# compare dfs: foo and merged_df3_com
all.equal(foo, merged_df3)
summary(comparedf(foo, merged_df3))
# sanity check
cat('Checking nrows in merged_df3_comp')
if(nrow(merged_df3_comp) == nrow(merged_df3)){
cat('NO NAs detected in merged_df3 in AF|OR cols'
,'\nNo. of rows are identical: ', nrow(merged_df3))
} else{
if(nrow(merged_df3_comp) == nrow(merged_df3) - na_count_df3) {
cat('PASS: NAs detected in merged_df3 in AF|OR cols'
, '\nNo. of NAs: ', na_count_df3
, '\nExpected no. of rows in merged_df3_comp: ', nrow(merged_df3) - na_count_df3
, '\nGot no. of rows: ', nrow(merged_df3_comp))
}
}
#=======================================================================
#*********************
# writing 1 file in the style of a loop: merged_df3
# print(output dir)
#i = 'merged_df3'
#out_filename = paste0(i, '.csv')
#outfile = paste0(outdir, '/', out_filename)
#cat('Writing output file: '
# ,'\nFilename: ', out_filename
# ,'\nPath: ', outdir)
#template: write.csv(merged_df3, 'merged_df3.csv')
#write.csv(get(i), outfile, row.names = FALSE)
#cat('Finished writing: ', outfile
# , '\nNo. of rows: ', nrow(get(i))
# , '\nNo. of cols: ', ncol(get(i)))
#%% write_output files; all 4 files:
outvars = c('merged_df2'
, 'merged_df3'
, 'merged_df2_comp'
, 'merged_df3_comp')
cat('Writing output files: '
, '\nPath:', outdir)
for (i in outvars){
# cat(i, '\n')
out_filename = paste0(i, '.csv')
# cat(out_filename, '\n')
# cat('getting value of variable: ', get(i))
outfile = paste0(outdir, '/', out_filename)
# cat('Full output path: ', outfile, '\n')
cat('Writing output file:'
,'\nFilename: ', out_filename,'\n')
write.csv(get(i), outfile, row.names = FALSE)
cat('Finished writing: ', outfile
, '\nNo. of rows: ', nrow(get(i))
, '\nNo. of cols: ', ncol(get(i)), '\n')
}
# alternate way to replace with implicit loop
# FIXME
#sapply(outvars, function(x, y) write.csv(get(outvars), paste0(outdir, '/', outvars, '.csv')))
#=======================================================================
#%% merging mut_pos_occurrence and sample_pos_occurence
# FIXME
#cat('Merging dfs with positional frequency from mcsm and meta_data'
# , '\nNcol in mut_pos_occurrence:', ncol(mut_pos_occurrence)
# , '\nncol in sample_pos_occurence:', ncol(sample_pos_occurrence)
# ,'\nlinking col:', intersect(colnames(sample_pos_occurrence), colnames(mut_pos_occurrence))
# ,'\nfilename: merged_df4')
#merged_df4 = merge(sample_pos_occurrence, mut_pos_occurrence
# , by = 'position'
# , all = T)
#out_filename4 = 'mut_and_sample_freq.csv'
#outfile4 = paste0(outdir, '/', out_filename4)
#*************************
# clear variables
rm(mcsm_data, meta_with_afor, foo, drug, gene, gene_match, indir, merged_muts_u, meta_muts_u, na_count, orig_col, outdir)
rm(mut_pos_occurrence, sample_pos_occurrence)
#rm(merged_df4)
#%% end of script
#=======================================================================