I have several gene-expression datasets and would like to perform a meta-analysis by combining p-values using Fisher's method. I used to perform this using p-values derived from differential expression analysis. I came then across a package named rf permute for R. In my understanding, this package is used to estimate significance (p-values) from computed variable importances (obtained by RandomForest) by permuting the response variable. Will I be able to use random Forest derived p-values for meta-analysis instead of those obtained from differential expression? Is this feasible?
Are there any original datasets like mean, mean differences with SD or SE? Or number of cases with endpoint? Just to make sure - you are asking if there is a method of combining just p values? I have never heard of such method.
Dear Marco, you may use STATA software for meta-analysis (it is a very powerful software for meta-analysis with various commands). "metap" command combines p-values by using Fisher’s method, Edgington’s additive method, or Edgington’s normal curve method (If you have a dataset which contains data for all studies, then the "byvar" command can be used to derive the p-values for the individual studies). It was released in 1999 as a version 6 command and was last updated in 2000. It requires the user to input a p-value for each study.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools