scripts/prepare_maxquant.R

In vladpetyuk/LewyBodies.SN.BottomUp2: Collection of bottom-up LC-MS/MS data for Lewy bodies project

resPath <- "../notinst/maxquant/Lewy2b/"


library("vp.misc")
m <- readMaxQuantProtGroups(resPath, quantType = 'LFQ intensity')

# METAINFO
clin <- read.delim("../notinst/final_matches_info.txt", as.is=TRUE, check.names=FALSE)
clin[['CSF available?']] <- NULL
# adding sample prep metadata. Batches are true except 
# with some differences for Lewy3
batch.assn <- read.delim("../notinst/bottom_up_sample_batch_assn.txt", as.is=TRUE, check.names=F)
colnames(batch.assn)[colnames(batch.assn) == 'SampleName'] <- 'sample.id'
meta <- merge(clin, batch.assn)
rownames(meta) <- meta$sample.id

# to.retain <- c("PlatePos", "PlateRow", "PlateCol", "HomogenizationBatch")
# to.retain <- c(to.retain,
#                c("projid","age_death","msex","educ","pmi","parksc_lv",
#                "henl_4gp","niareagansc","tang_dens_sn","amy_mean_sn"))
# meta <- meta[,to.retain]
subject.type <- sub("\\d+(C.*)","\\1",rownames(meta))
subject.type <- ifelse(subject.type == "Cs", "case",
                       ifelse(subject.type == "Ct1", "control.1", "control.2"))
meta$subject.type <- as.factor(subject.type)
match.group <- sub("(\\d+)C.*","\\1",rownames(meta))
meta$match.group <- as.factor(sprintf("%02d", as.numeric(match.group)))
# NOTE, SOME INDEPENDENT (PHENOTYPE OR TECHNICAL) VARIABLES ARE FACTORS
meta$PlateCol <- as.factor(meta$PlateCol)
meta$PlateRow <- as.factor(meta$PlateRow)
pData(m) <-cbind(pData(m), meta[sampleNames(m),])



# DMS data
dms <- file.path("../notinst", 
                 paste(tolower(sub(".*(Lewy\\d).*","\\1", resPath)), 
                       "_dms_data.txt", sep=''))
dms <- read.delim(dms, as.is=TRUE, check.names=FALSE)
library("lubridate")
o <- order(parse_date_time(dms[["Acq. End"]], "%m%d%y %I%M %p"))
dms <- dms[o,c("Dataset","Experiment","Acq. End")]
dms <- subset(dms, Dataset %in% pData(m)$dataset.name)
rownames(dms) <- sub("Lewy\\d?_", "", dms$Experiment)
dms$Experiment <- NULL
colnames(dms) <- c("Dataset","Acq.End")
dms$run.order <- 1:nrow(dms)
pData(m) <- cbind(pData(m), dms[sampleNames(m),])



# now a bit PRE-PROCESSING
# 1. log2-transform
# 2. zero-center
# 3. Remove proteins with low observation count
exprs(m) <- log2(exprs(m))
exprs(m) <- sweep(exprs(m), 1, rowMeans(exprs(m), na.rm=TRUE), '-')
m <- m[rowSums(!is.na(exprs(m))) > 1,]



# SAVING
save(m, file=sprintf("../data/%s_maxquant.RData", 
                     basename(resPath)),
     compress='xz')