search

tleonardi / nanocompore

RNA modifications detection from Nanopore dRNA-Seq data
https://nanocompore.rna.rocks
GNU General Public License v3.0
78 stars 12 forks source link

p values corrected for multi-testing #124

Closed Huanle closed 4 years ago

Huanle commented 4 years ago

Hi @tleonardi , Are the p-values in the results of nanocompore already corrected for multi-testing? From the manuscript, it seems to be the case for those values computed by taking into sequence account. But is it the same for single-base senario? Thanks a lot.

a-slide commented 4 years ago

Hi @Huanle The pValues in the final text report should be adjusted for multiple tests. I am not sure I understand what you mean by single-base senario. Ad

Huanle commented 4 years ago

Hi @a-slide ,

thanks a lot for yout quick reply.

in the manuscript it says 'Thus, to evaluate the effect of modifications on the proximal sequence context, we implemented an established statistical method that combines the non-independent p-values to produce a combined p-value representative of multiple neighbouring kmers, The p-values are then corrected for multiple tests using Benjamini-Hochberg’s procedure ....‘. By single-base, i mean p-values computed for the modification site wtithout accounting for its sequence context (--sequence_context 0).

tleonardi commented 4 years ago

Hi @Huanle, yes, all p-values in the output files (txt, bed and bedg) are corrected with the BH procedure. This applies both to p-values deriving from the combination of adjacent kmers (_context columns in the output file) as well as to the p-values of individual kmers. The raw p-values (i.e. prior to BH correction) are stored into the SampCompDB object, if you are interested I can provide info on how to access those.