filter-mass_dataset: filter_samples

In tidymass/massdataset: massdataset

filter_samples

Description

Filter samples/variables based on the conditions

Usage

filter_samples(object, flist, prune = TRUE, apply_to = "all")

filter_variables(
  object,
  flist,
  prune = TRUE,
  apply_to = "all",
  according_to_samples = "all"
)

Arguments

object	(required) mass_dataset class object.
flist	(required) A function or list of functions that take a vector of abundance values and return a logical.
prune	(optional) A logical. Default FALSE. If TRUE, then the function returns the pruned mass_dataset-class object, rather than the logical vector of variables that passed the filter.
apply_to	(required) what variables you want to apply this function. Default is "all". If you only want to apply to specific variables, please set it as a vector of sample names. Other variables will be set as TRUE.
according_to_samples	(required) What samples used to filter variables. Default is "all". If you want to use only several samples, provide they names as a vector.

Value

A logical vector equal to the number of samples/variables in mass_dataset-class. Alternatively, if prune==TRUE, the pruned mass_dataset-class object is returned instead.

Author(s)

Xiaotao Shen [e-mail]

Examples

data("expression_data")
data("sample_info")
data("variable_info")

object =
  create_mass_dataset(
    expression_data = expression_data,
    sample_info = sample_info,
    variable_info = variable_info,
  )
 
filter_samples(object, function(x) {
  sum(is.na(x)) / length(x) < 0.4
})

filter_samples(object, function(x) {
  sum(is.na(x)) / length(x) < 0.4
}, prune = FALSE)

##only apply to Subject sample
object2 =
filter_samples(
 object = object,
 flist = function(x) {
   sum(is.na(x))/length(x) < 0.2
 },
 prune = TRUE,
 apply_to = get_sample_id(object)[extract_sample_info(object)$class == "Subject"]
)

object2
library(tidyverse)
data("expression_data")
data("sample_info")
data("variable_info")

object =
  create_mass_dataset(
    expression_data = expression_data,
    sample_info = sample_info,
    variable_info = variable_info,
  )
object

####Filter variables which have more than 50% MVs in all samples.
library(tidyverse)
filter_variables(object, function(x) {
  sum(is.na(x)) / length(x) < 0.5
}, prune = FALSE) %>%
  head()

filter_variables(object, function(x) {
  sum(is.na(x)) / length(x) < 0.5
},
prune = TRUE)

####Filter variables which have more than 50% MVs in only QC samples.
filter_variables(
  object,
  flist = function(x) {
    sum(is.na(x)) / length(x) < 0.5
  },
  prune = TRUE,
  according_to_samples =
    get_sample_id(object)[extract_sample_info(object)$class == "QC"]
)

####Filter variables which have more than 50% MVs in QC or subject samples.
idx1 =
  filter_variables(
    object,
    flist = function(x) {
      sum(is.na(x)) / length(x) < 0.5
    },
    prune = FALSE,
    according_to_samples =
      get_sample_id(object)[extract_sample_info(object)$class == "QC"]
  )

idx2 =
  filter_variables(
    object,
    flist = function(x) {
      sum(is.na(x)) / length(x) < 0.5
    },
    prune = FALSE,
    according_to_samples =
      get_sample_id(object)[extract_sample_info(object)$class == "Subject"]
  )

idx =
  which(idx1 | idx2)

object2 = object[idx,]

object2

####filter variables with RSD (in QC samples) < 30
object3 =
filter_variables(
  object = object,
  flist = function(x) {
    rsd = sd(x) * 100 / mean(x)
    rsd = ifelse(is.na(rsd), 100, rsd)
    rsd < 30
  },
  apply_to = "all",
  prune = TRUE,
  according_to_samples = get_sample_id(object)[extract_sample_info(object)$class == "QC"]
)

object3

tidymass/massdataset documentation built on Sept. 14, 2024, 4:41 p.m.