inst/dataprep/ID93.R

In MiguelRodo/VaccCompData: Data package for novel TBV vaccine comparison paper

# Loading -----------------------------------------------------------------

# names of current objects
currObjVec = ls()

# find library containing raw data
pkgPath = system.file( package = "VaccCompData" )

# libraries
source( file = paste0( pkgPath, "/dataprep/libraries.R") )

#read in data
fileLoc = paste0( pkgPath, "/extdata/id93.xlsx" )
idriTbl = as_tibble( read.xlsx( fileLoc, sheet = 'Sheet 1') )

# Data Prep ---------------------------------------------------------------

# na check

# table of NA or not
naTbl = sapply( idriTbl[,8:ncol(idriTbl) ], Vectorize(is.na) ) %>% as_tibble()

# columns with NAs
naTbl %>%
  summarise_all( function(x) sum(x*1 ) ) %>%
  gather( key = cytCombo, value = naCount, `CD8+/G+T+2+17+`:`CD4+/G-T-2-17+` ) %>%
  filter( naCount > 0 ) # only for that one triple positive. # it's for a PHA anyway.

# unique values

unique( idriTbl$SAMPLE.ID ) # "ID93+GLA/SE" "Placebo". Must remove Placebos
unique( idriTbl$QFT.Status ) # "QFT-" "QFT+". Correct, but convert to binary
unique( idriTbl$TUBE.NAME ) # "PHA" "Rv1813" "Rv2608" "Rv3619" "Rv3620" "UNS". Correct.
unique( idriTbl$SAMPLE.ID ) # all numbers, fine.
unique( idriTbl$COHORT ) # 1 2 3 4. Use group 1 and 3.
unique( idriTbl$Study.Day) #  0  14  42 112 126 196 294. Fine.

# filter rows and remove unnecessary columns
idriTbl %<>%
  filter( COHORT == 1 | COHORT == 3 ) %>% # chose the two cohorts that give the same id93 and gla-se dose, but to different infection statuses
  filter( RANDOMISATION == 'ID93+GLA/SE' ) %>%
  `[`( -c( 2, 5 ) )
unique( idriTbl$SAMPLE.ID)
firstCytCol = 6

# rename columns and edit rows
idriTbl %<>%
  rename( infxn = QFT.Status, stim = TUBE.NAME, ptid = SAMPLE.ID, timePoint = Study.Day,
          group = COHORT ) %>%
  mutate( infxn = sapply( infxn, function(x) ifelse( x == 'QFT-', "0", "1" ) ),
          stim = str_to_lower( stim ),
          ptid = as.character( ptid ),
          timePoint = as.character( timePoint ) )

#  remove pha stim
idriTbl %<>%
  filter( stim != 'pha' )

idriTbl %>%
  group_by( group ) %>%
  summarise( n_distinct( infxn ) )

### BOOLEAN COLUMN NAMES

# remove useless columns
#idriTbl %<>% select( -c( `CD8+.Total.cytokine.response`, `CD4+.TOTAL` ))

# get actual names
cytNameVec = str_to_upper( tbl_vars( idriTbl )[firstCytCol:ncol( idriTbl )] )

# get replacement names
for (i in 1:length(cytNameVec)){ #loop over names
  currCyt = cytNameVec[i]
  #check if we have a positive in a given location
  #if so, add a p onto a string. If not, add an n
  locVec = c( 7, 9, 11, 14 )
  signVec = str_sub(currCyt, locVec,locVec)
  for (j in 1:length(signVec)){
    if (signVec[j] == '+'){
      str_sub(currCyt, locVec[j], locVec[j]) = 'p'
    } else if (signVec[j] == '-'){
      str_sub(currCyt, locVec[j], locVec[j]) = 'n'
    }
  }
  cytNameVec[i] = currCyt
}

# replace names

origCytNameVec = tbl_vars(idriTbl)[firstCytCol:ncol(idriTbl)]
colNameCodeDF = data_frame(x = str_c("idriTbl %<>% rename( `", cytNameVec, "` = `", origCytNameVec, "`)"))

for( i in 1:nrow( colNameCodeDF ) ) eval(parse(text = colNameCodeDF[i,1]))

### RE-ORDER

idriTbl = changeCytOrder( dataTibble = idriTbl, firstCytColIndex = firstCytCol,
                           firstCytNameLoc = 6, signLocVec = c( 7, 9, 11, 14 ),
                           orderLocVec = c( 1, 3, 2, 4 ), cdPresent = TRUE )
# arrange rows

idriTbl %<>% arrange( group, infxn, ptid, timePoint, stim )

# Data Checking -----------------------------------------------------------

# stim

# number of stims
idriTbl %>%
  group_by( infxn, group, ptid, timePoint ) %>%
  summarise( stimCount = n() ) %>%
  filter( stimCount != 5 ) # all have five stims.

# number of unique stims
idriTbl %>%
  group_by( infxn, group, ptid, timePoint ) %>%
  summarise( uniqueStimCount = n_distinct( stim ) ) %>%
  filter( uniqueStimCount != 5 ) # all have five unique stims.

# Subtraction -------------------------------------------------------------

###########################
### SUB UNS FROM OTHERS ###
###########################

# merge
idriTbl %<>% arrange( group, infxn, ptid, timePoint, stim )

# subtraction
idriITbl = idriTbl %>%
  group_by( group, infxn, ptid, timePoint ) %>%
  mutate_if( is.numeric, function(x) x - x[5] ) %>%
  ungroup() %>%
  filter( stim != "uns" )


# summation
idriTbl %<>%
  group_by( group, infxn, ptid, timePoint ) %>%
  summarise_if( is.numeric, function(x) sum( x[ 1:4 ] ) - 4 * x[ 5 ] ) %>%
  ungroup()

# Final Editing -----------------------------------------------------------

#idriITbl %<>% mutate( ptid = str_c( "a", ptid, sep = "" ) )

#idriTbl %<>% mutate( ptid = str_c( "a", ptid, sep = "" ) )


# Testing -----------------------------------------------------------------

idriTbl %>% filter( ptid %in% c("id93_1", "id93_7", "id93_32", "id93_43" ) ) %>% group_by( ptid ) %>% slice(1) %>% select( group, ptid )
 idriITbl
testRow1 = idriITbl %>% filter( ptid == 'id93_1' & timePoint == 0 & stim == "rv3619" )
if( testRow1$CD4Gp2nTn17n != (0.009664287 - 0.007563361) ) stop( "idriITbl error" )
testRow2 = idriITbl %>% filter( ptid == 'id93_43' & timePoint == 294 & stim == "rv2608" )
if( testRow2$CD8Gp2nTn17n != (0.00115 - 0.000528) ) stop( "idriITbl error" )
testRow3 = idriITbl %>% filter( ptid == 'id93_7' & timePoint == 112 & stim == "rv3620" )
if( testRow3$CD8Gp2nTp17n != (0.1967563 - 0.1723022) ) stop( "idriITbl error" )

testRow1 = idriTbl %>% filter( ptid == 'id93_32' & timePoint == 126 )
if( signif( testRow1$CD4Gn2nTp17n, 7 ) != signif( (0.005361815 + 0.01608234 + 0.01048621 + 0.02476414 - 4 * 0.008231904), 7 ) ) stop( "idriTbl error" )

# Clear workspace ---------------------------------------------------------

# remove all objects but initial objects (currObjVec) and processed data tibble (idriTbl)
rm( list = setdiff( ls(), c( currObjVec, "idriTbl", "idriITbl" ) ) )