Correction of meta-q value

[Ivo Kwee]

I have an additional question. Omics Playground reports the meta q-value which is the aggregate of the used method results. So correction for multiple testing occurs ath the individual test results? Which correction is used, Bonferroni or Benjamini-Hochberg? Or is the meta-q corrected?

[31.08.2021 Thorben S., via email]

[BigOmics Analytics Team]

Great question. In Omics Playground, the multiple test correction is done before q-value aggregation. So each method has it's own independent p-value and adjusted p-value. Aggregation of q-value (meta.q) is done after adjustment of individual p-values. True it may be possible to adjust the aggregated p-value for multiple test correction. Not sure if this is preferred. On one hand it could correct for the multiplicity of "multiple methods", on the other hand perhaps each method may not be equally uniform and exhibit different p-value distribution. By doing the aggregation after multiple test correction, by taking the maximum q-value, the interpretation is more straightforward: all methods agree. But I would be interested if there is a comparison of the methods either in literature or a comparison to be done.

Ivo