summary_stats

This function provides the following summary statistics:

  1. median_survival: Median survival with 95% confidence intervals

  2. logrank: Log-rank test - is there a difference between two (or more) survival curves?

  3. pairwise: Pairwise log-rank test with p-value correction between conditions. Default correction: "BH".

summary_stats(df_isurv, type = "all", p_adjust_method = "BH")

Arguments

df_isurv

a data.frame of survival data, with one individual per row. If starting with bulk survival data (multiple observations per row), run get_indiv_surv() first.

type

character, either "median_survival", "logrank", "pairwise" or "all". Default: "all"

p_adjust_method

character, either "holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none". For details, see ?stats::p.adjust.

Value

If type == "all": a list object containing results from statistical tests. Otherwise, returns a ggplot object containing the survival/mortality curve

Details

  • logrank: Calls survival::survdiff() to test if there is a difference between two or more survival curves using the G^rho family of tests

  • pairwise: Calls survminer::pairwise_survdiff(), which calculates pairwise comparisons between groups with corrections for multiple testing.

Examples

df_isurv <- get_indiv_surv(sample_data,
                           sample_order = c("WT", "Drug1", "Drug2"))
#> Joining with `by = join_by(x, condition, day, sex, status)`

stats <- summary_stats(df_isurv)
#> 
#> call: formula = Surv(day, status) ~ condition
stats
#> $median_survival
#> Call: survfit(formula = Surv(day, status) ~ condition, data = df_isurv)
#> 
#>                  n events median 0.95LCL 0.95UCL
#> condition=WT    50     47   20.0      19      21
#> condition=Drug1 50     46    4.5       4       6
#> condition=Drug2 50     42   33.0      32      35
#> 
#> $logrank
#> Call:
#> survival::survdiff(formula = Surv(day, status) ~ condition, data = df_isurv)
#> 
#>                  N Observed Expected (O-E)^2/E (O-E)^2/V
#> condition=WT    50       47     31.2      7.94      12.9
#> condition=Drug1 50       46     12.5     89.99     128.1
#> condition=Drug2 50       42     91.3     26.60     123.3
#> 
#>  Chisq= 186  on 2 degrees of freedom, p= <2e-16 
#> 
#> $pairwise
#> 
#> 	Pairwise comparisons using Log-Rank test 
#> 
#> data:  df_isurv and condition 
#> 
#>       WT      Drug1  
#> Drug1 2.3e-14 -      
#> Drug2 < 2e-16 < 2e-16
#> 
#> P value adjustment method: BH 
#>