Perform bootstrap estimation of confidence intervals of ranked HR scores

estimate carries out estimation of bootstrap confidence intervals on ranked score data. Returns a hrest object of the result Proceeeds by calculating score ranks, then bootstrapping ranks in non-control groups retaining the mean for each bootstrap iteration. Calculates low and high quantiles of bootstrap mean distributions for each group. If technical replicates are provided in a second grouping column these will be averaged before proceeding.

Usage

estimate(df, ..., control = "A", nits = 100, low = 0.025, high = 0.975)

Arguments

df: data frame of score and group data. Contains minimally a score and group column
...: bare names of columns to use, minimally the score column and the group column in that order. Optionally a third technical replicate column can be provided
control: the value of the grouping column taken to be the control group
nits: the number of bootstap iterations to be done
low: the low probability value of the quantile
high: the high probability value of the quantile

Value

a list object of class "hrest"

Examples


 d1 <- make_data()
 estimate(d1, score, group)
#> besthr (HR Rank Score Analysis with Bootstrap Estimation)
#> =========================================================
#> 
#> Control: A
#> 
#> Unpaired mean rank difference of A (14.6, n=10) minus B (6.4, n=10)
#>  8.2
#> Confidence Intervals (0.025, 0.975)
#>  3.82375, 9.2575
#> 
#> 100 bootstrap resamples.

 d2 <- make_data2()
 estimate(d2, score_column_name, sample_column_name, rep_column_name )
#> besthr (HR Rank Score Analysis with Bootstrap Estimation)
#> =========================================================
#> 
#> Control: A
#> 
#> Unpaired mean rank difference of A (5, n=3) minus B (2, n=3)
#>  3
#> Confidence Intervals (0.025, 0.975)
#>  1, 3
#> 
#> 100 bootstrap resamples.

# \donttest{
 d3 <- make_data3()
 estimate(d3, score, sample, rep, nits = 1000)
#> besthr (HR Rank Score Analysis with Bootstrap Estimation)
#> =========================================================
#> 
#> Control: A
#> 
#> Unpaired mean rank difference of A (5, n=3) minus B (2, n=3)
#>  3
#> Confidence Intervals (0.025, 0.975)
#>  1, 3
#> 
#> Unpaired mean rank difference of A (5, n=3) minus C (8, n=3)
#>  -3
#> Confidence Intervals (0.025, 0.975)
#>  7, 9
#> 
#> 1000 bootstrap resamples.
# }