cellstudy1202.Rmd

---
title: "xcms tutorial"
author: "Chuyao Wang"
date: "`r Sys.Date()`"
output: 
  html_document: 
    number_sections: true # automatic section numbering
    toc: yes # table of content
    toc_depth: 3
    toc_float: true
    smooth_scroll: true
    toc_collapsed: true
    fig_caption: yes
    df_print: kable
knit: (function(inputFile, encoding) {
  rmarkdown::render(inputFile, encoding = encoding, output_dir = "./html") })
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)

library(dplyr)
library(stringr)
library(openxlsx)
```

Load data and clean up colnames.
```{r}
data <- read.xlsx("./data/cellstudy20221202_CD33_result.xlsx")
data <- data %>%
  `[`(,!str_detect(colnames(.),"QC")) %>%
  `[`(,!str_detect(colnames(.),"methanol"))

colnames(data) <- colnames(data) %>% str_remove(.,"768O") %>% str_remove(.,"_Waters_Zhu_ddMS2_[:digit:]+_[:digit:]+") %>% str_remove(.,".raw.") %>% str_remove("\\(F[:digit:]+\\)") %>% str_remove("._") %>% str_replace(":","_")
```

```{r}
# Get max area
featureData <- select(data,colnames(data)[!str_detect(colnames(data),"Area")&!str_detect(colnames(data),"Peak")])
areas <- select(data,colnames(data)[str_detect(colnames(data),"Area")])
areas.new <- lapply(1:(ncol(areas)/4),function(i) {
  idx <- (i-1)*4+c(1:4)
  return(apply(areas[,idx],MARGIN=1,max,na.rm=TRUE))
}) %>%
  data.frame()
colnames(areas.new) <- colnames(areas) %>% str_remove(.,"_[:letter:]+_[:digit:]") %>% unique

data.new <- cbind(featureData,areas.new)

# Remove duplicate metabolites
# If there are duplicates, merge by taking the max of each metabolite. Then keep only one row
duped <- data.new$Formula %>% table %>% as.data.frame() %>% `[`(.$Freq>1,) %>% `[`(,1) %>% as.vector()
for (i in duped) {
  idx <- which(data.new$Formula %in% i)
  newrow <- data.new[c(idx),] %>%
    select(colnames(data.new)[str_detect(colnames(data.new),"Area")]) %>%
    sapply(.,max)
  for (j in idx) {
    data.new[j,which(str_detect(colnames(data.new),"Area"))] <- newrow
  }
  data.new$Name[idx[-1]] <- "dup"
}
data.new <- data.new %>% filter((Name != "dup")|(is.na(Name)))

# Give NA names an ID
data.new$Name[which(is.na(data.new$Name))] <- paste("Unknown",seq_along(which(is.na(data.new$Name))),sep="_")
```

```{r}
# Transpose and process
# data.new <- data.new[which(!str_detect(data.new$Name,"Unknown")),] # Removes unknown compounds
areas.out <- data.new %>% select(colnames(data.new)[str_detect(colnames(data.new),"Area")]) %>% t %>% as.data.frame
colnames(areas.out) <- data.new$Name
areas.out <- mutate(areas.out, grp = rownames(areas.out) %>% str_split_fixed(.,"_",3) %>% `[`(,3)) %>%
  select(grp,everything())
```