SAGx (version 1.46.0)

rank.trend: Trend analysis based on ranks

Description

Ranks are used to score genes with respect to degree of agreement to a given trend or pattern, Lehmann (1974) p.294.

Usage

rank.trend(data = x, pattern = c(1:ncol(data)), har = FALSE)

Arguments

data

A data frame with one array in each column

pattern

A permutation of the integers 1:ncol(data)

har

logical parameter indicating whether or not a score based on Hardy's theorem shall be calculated.

Value

• A list with the components
• score
  the rank score for each gene
• hardy
  if har = TRUE the hardy score, NULL otherwise
• pvals
  the p-values for the null hypothesis of no trend

Details

The rank scores gives a higher weight to a deviation from trend in more distant obseveations than a deviation between neighbouring observations. The p-values are calculated through a normal approximation.

References

Lehmann, E.L. (1975) Nonparametrics: Statistical Methods Based on Ranks, Holden-Day

Examples

# not run
D <- c(123, 334, 578, 762, 755, 890)
rank.trend(data = t(as.matrix(D)), har = TRUE)
#     Trend score Hardy score p-value for no trend
# [1,]           2          90           0.01750284