read_from_excel: Read formatted Excel files

In antonvsdata/notame: Workflow for non-targeted LC-MS metabolic profiling

Read formatted Excel files

Description

Reads data from an Excel file of the following format:

• Left side of the sheet contains information about the features, size features x feature info columns

• Top part contains sample information, size sample info variables x samples

• The middle contains the actual abundances, size features x samples

See the vignette for more information. This function separates the three parts from the file, and returns them in a list

Usage

read_from_excel(
  file,
  sheet = 1,
  id_column = NULL,
  corner_row = NULL,
  corner_column = NULL,
  id_prefix = "ID_",
  split_by = NULL,
  name = NULL,
  mz_limits = c(10, 2000),
  rt_limits = c(0, 20),
  skip_checks = FALSE
)

Arguments

file	path to the Excel file
sheet	the sheet number or name
id_column	character, column name for unique identification of samples
corner_row	integer, the bottom row of sample information, usually contains data file names and feature info column names. If set to NULL, will be detected automatically.
corner_column	integer or character, the corresponding column number or the column name (letter) in Excel. If set to NULL, will be detected automatically.
id_prefix	character, prefix for autogenerated sample IDs, see Details
split_by	character vector, in the case where all the modes are in the same Excel file, the column names of feature data used to separate the modes (usually Mode and Column)
name	in the case where the Excel file only contains one mode, the name of the mode, such as "Hilic_neg"
mz_limits	numeric vector of two, all m/z values should be in between these
rt_limits	numeric vector of two, all retention time values should be in between these
skip_checks	logical: skip checking data integrity. Not recommended, but sometimes useful when you just want to read the data in as is and fix errors later. NOTE: Sample_ID and QC columns will not be constructed. The data integrity checks need to be passed when contstructing MetaboSet objects.

Details

Only specify one of split_by and name. The feature data returned will contain a column named "Split", which is used to separate features from different modes. Unless a column named "Feature_ID" is found in the file, a feature ID will be generated based on the value of "Split", mass and retention time. The function will try to find columns for mass and retention time by looking at a few common alternatives, and throw an error if no matching column is found. Sample information needs to contain a row called "Injection_order", and the values need to be unique. In addition, a possible sample identifier row needs to be named "Sample_ID", or to be specified in id_column, and the values need to be unique, with an exception of QC samples: if there are any "QC" identifiers, they will be replaced with "QC_1", "QC_2" and so on. If a "Sample_ID" row is not found, it will be created using the id_prefix and injection order.

Value

list of three data frames:

• exprs: the actual abundances, size features x samples

• pheno_data: sample information, size sample info variables x samples

• feature_data: information about the features, size features x feature info columns

antonvsdata/notame documentation built on Sept. 14, 2024, 11:09 p.m.