R 中的双向 ANOVA 后每组的多重比较（事后）

Question

I'm very new to R and need some help getting it to do the analysis I want. I have a dataframe like this (simplified):

   Gene time      Expression
1    Gene1       W1  18.780294
2    Gene2       W1  13.932397
3    Gene3       W1  20.877093
4    Gene1       W2   9.291295
5    Gene2       W2  10.939570
6    Gene3       W2  12.236713
7    Gene1       W3  13.810722
8    Gene2       W3  23.944473
9    Gene3       W3  17.355429

I want to know if there is a significant difference between the mean 'Expression' values for the Genes (Gene 1, 2, 3,) at each time point (Week1, W2, W3...). For example, I'm interested to know if at week 3( W3) the expression of Gene1 is significantly different than Gene2 I have performed an ANOVA and TukeyHSD

my_ANOVA = aov(Expression ~ Gene , data = my_Data)
summary(my_ANOVA)

my_posthoc =TukeyHSD(my_ANOVA, which = "Gene") 
my_posthoc

the results are

  Tukey multiple comparisons of means
    95% family-wise confidence level

Fit: aov(formula = Expression ~ Gene, data = my_Data)

$Gene
                 diff       lwr      upr     p adj
Gene2-Gene1 2.3113763 -11.24096 15.86371 0.8631245
Gene3-Gene1 2.8623080 -10.69002 16.41464 0.8002795
Gene3-Gene2 0.5509317 -13.00140 14.10326 0.9914716

It gives me the total comparission of Gene1 to Gene2 for entire timeline. But I want to know specifically if in time W1, Gene1 and 2 are diffrerent, or what about in time W3

Answer 1

If you just want to split the data by time and run each analysis separately, that is pretty straightforward although the data set is too small to produce useable results:

my_Data.W <- split(my_Data, my_Data$time)
lapply(1:3, function(x) summary(aov(Expression~Gene, data=my_Data.W[[x]])))
lapply(1:3, function(x) TukeyHSD(aov(Expression~Gene, data=my_Data.W[[x]])))