rank.trend {SAGx} | R Documentation |
Ranks are used to score genes with respect to degree of agreement to a given trend or pattern, Lehmann (1974) p.294.
rank.trend(data = x, pattern = c(1:ncol(data)), har = FALSE)
data |
A data frame with one array in each column |
pattern |
|
har |
logical parameter indicating whether or not a score based on Hardy's theorem shall be calculated. |
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.
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 |
Per Broberg
Lehmann, E.L. (1975) Nonparametrics: Statistical Methods Based on Ranks, Holden-Day
# 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