In jaspershen/metflow2: Comprehensitive tools for metabolomics data processing and data analysis

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  # fig.width = 7, 
  # fig.height = 5,
  warning = FALSE,
  message = TRUE,
  out.width = "100%"
)

---

Introduction

metflow2 can used to processed the raw MS data for peak detection and alignment, and generate a peak table for next analysis.

Data preparation

The MS raw data from Mass Spectrometry should be converted to mzXML format and then placed in different folders according to their class, for example Subject, QC, and Blank samples.

Here we use the demo data from demoData packages. So please install demoData package first.

devtools::install_github("jaspershen/demoData")

You can convert raw MS data to mzXML format using ProteoWizard software. And the parameter setting is shown the below figure:

Data organization

All the mzXML format files should be placed in different folder according to sample type, such as QC, Subject, and Blank.

Run processData() function

We use the demo data from demoData to show how to process raw MS data using metflow2.

Load demo data

First we load the demo data from demoData package and then place them in a example folder.

library(demoData)
library(metflow2)
library(tidyverse)
library(xcms)
library(MSnbase)
library(mzR)

##create a folder named as `example`
path <- file.path(".", "example")
dir.create(path = path, showWarnings = FALSE)

##get demo data
mzxml_data <- system.file("mzxml", package = "demoData")

file.copy(from = file.path(mzxml_data, "POS"), 
          to = path, overwrite = TRUE, recursive = TRUE)

Now the demo mzXML data is in the ./example/POS folder. There are two folder in this directory, namely Subject and QC.

Data processing

Next, we use the process_data() function for peak detection and alignment.

new_path <- file.path(path, "POS")
setwd(path)
metflow2::process_data(
  path = ".",
  polarity = "positive",
  ppm = 15,
  peakwidth = c(5, 30),
  snthresh = 5,
  noise = 500,
  threads = 6,
  output.tic = TRUE,
  output.bpc = TRUE,
  output.rt.correction.plot = TRUE,
  min.fraction = 0.5,
  fill.peaks = FALSE,
  is.table = "is.xlsx",
  group.for.figure = "QC"
)

Some important arguments:

• ppm: Peak detection ppm. See the xcms package.

• peakwidth: Peak width. It is dependent on your column and LC system. See the xcms package.

• snthresh: Singal to noise threshold. See the xcms package.

• noise: Noise cutiff. See the xcms package.

• threads: The core number for performing.

• output.tic, output.bpc, and group.for.figure: Do you want to output the TIC or BPC of samples? Some times there are a lot of samples, so you can set the group.for.figure as the group name, for example, you can set it as QC.

• min.fraction: If one peak appears more than min.fraction sample in at least one group samples it will be kept.

• fill.peaks: Do you want to fill the missing values (NA)?

• is.table: If you add internal standards in your samples, you can provide the the is.table in the folder which your mzXML format data in. It must be xlsx format like the below figure shows:


Ohter parameters you can find here: processData().

Output results

After all done, all the results are placed in a new folder named as Result.

 * Peak_table.csv: is the peak table for next analysis.


• Peak_table_for_cleaning.csv: Some non-useful columns is deleted from the Peak_table.csv, this can be directory used for next data cleaning.


• intermediate_data: This is a folder which contains some intermediate data. If you want to re-run the data processing, please delete this folder first.

• feature_EIC: If you provide the internal standard table, the EICs of all the ISs will be placed in this folder.

From the EICs of the internal standards, you can access the quality of data.

jaspershen/metflow2 documentation built on Aug. 15, 2021, 4:38 p.m.