Filter protein data by proportion of quantified samples in group
filter_proteins_by_proportion.RdThis function is used to remove proteins from a DAList, filtering out proteins
based on levels of missing data in the "data" data frame of the DAList. The
grouping_column must be a column in the metadata of the DAList which lists the
group membership for each sample. Proteins must have a non-missing intensity
value in at least min_prop of samples within each group in order to be retained.
When min_prop leads to a non-integer value for a given group, it is rounded up:
e.g., with 10 samples in a group and a min_prop of 0.75, a protein must be present in
at least 8 samples to be retained. This function assumes that all missing values are encoded
as NA. See zero_to_missing and missing_to_zero for helper
functions to convert missing values to and from 0.
Arguments
- DAList
A DAList object to be filtered.
- min_prop
The minimum proportion of samples within a group in which a protein must be found for it to be retained. Must be a numeric value from 0 to 1.
- grouping_column
The name of the column in the metadata which provides the group membership for each sample. Default is "group".
Details
filter_proteins_by_proportion() is useful when sample sizes are not constant
across groups, allowing similar levels of missingness across groups even when sample sizes
are imbalanced.
Examples
if (FALSE) { # \dontrun{
# Suppose the DAList contains data from 40 samples across 2
# experimental groups, with 15 samples in one group and 25 in the other.
# and group membership listed in a column names "group"
# Strict filtering:
# no missing data
# Proteins must be present in all samples in all groups
filtered <- filter_proteins_by_proportion(DAList,
min_prop = 1,
grouping_column = "group")
# moderate filtering:
# protein must be present in at 50% of samples within each group.
# That is, 8 samples in the group of 15 and 13 samples in the group of 25.
filtered <- filter_proteins_by_proportion(DAList,
min_prop = 0.5,
grouping_column = "group")
# Filtering functions can be chained together
filtered <- DAList |>
filter_proteins_by_annotation(!grepl(pattern = "keratin",
x = protein_name)) |>
filter_proteins_by_proportion(min_prop = 0.5,
grouping_column = "group")
} # }