tna (version 1.3.1)

compare_sequences: Compare Sequences Between Groups

Description

Performs comprehensive sequence comparison analysis between groups. All patterns of the sequences (subsequences of specific length) are extracted from all sequences in each group. The pattern frequencies are compared between the groups using a permutation test. The reported effect size is the difference between the observed test statistic (sum of squared differences between the observed and expected counts) and the mean value over the permutation samples divided by their standard deviation times square root of the number of observations.

Usage

compare_sequences(x, ...)
# S3 method for default
compare_sequences(
  x,
  group,
  sub,
  min_freq = 5L,
  test = TRUE,
  iter = 1000L,
  adjust = "bonferroni",
  ...
)
# S3 method for group_tna
compare_sequences(
  x,
  sub,
  min_freq = 5L,
  test = TRUE,
  iter = 1000L,
  adjust = "bonferroni",
  ...
)

Value

A tna_sequence_comparison object, which is a data.frame with columns giving the names of the patterns, pattern frequencies, pattern proportions (within patterns of the same length), effect sizes, and p-values of the tests.

Arguments

x

A group_tna object or a data.frame containing sequence data in wide format.

...

Not used.

group

A vector indicating the group assignment of each row of the data/sequence. Must have the same length as the number of rows/sequences of x. Alternatively, a single character string giving the column name of the data that defines the group when x is a wide format data.frame or a tna_data object.

sub

An integer vector of pattern lengths to analyze. The default is 2:5.

min_freq

An integer giving the minimum number of times that a specific pattern has to be observed in each group to be included in the analysis. The default is 5.

test

A logical value indicating whether to test the differences of pattern counts between the groups using a permutation test. The default is TRUE.

iter

An integer giving the number of iterations for the permutation test. The default is 1000.

adjust

A character string naming the multiple comparison correction method (default: "bonferroni"). Supports all stats::p.adjust methods: "holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none". The adjustment is carried out within sequences of the same length.

See Also

Model comparison functions compare(), compare.group_tna(), plot.tna_comparison(), plot.tna_sequence_comparison(), plot_compare(), plot_compare.group_tna(), print.tna_comparison(), print.tna_sequence_comparison()

Examples

# \donttest{
# Use a subset to keep the example fast.
idx <- c(1:500, 1001:1500)
group <- c(rep("High", 500), rep("Low", 500))
comp <- compare_sequences(group_regulation[idx, ], group)

# With permutation test (small number of iterations for CRAN)
comp_test <- compare_sequences(
  group_regulation[idx, ],
  group,
  test = TRUE,
  iter = 5
)
# }