search

tleonardi / nanocompore

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

Where is the fold change value #204

Closed JeremyQuo closed 1 year ago

JeremyQuo commented 1 year ago

The result tsv file I run from SampComp is like this,

pos chr genomicPos ref_id strand ref_kmer GMM_logit_pvalue KS_dwell_pvalue KS_intensity_pvalue GMM_cov_type GMM_n_clust cluster_counts Logit_LOR 10 NA NA NC_004162.2 NA AGACA 0.3297790012387313 0.4134951318088154 0.00393641138422188 full 2 WT_1:22/26__IVT_1:1035/750 -0.48205942113047806 11 NA NA NC_004162.2 NA GACAC 0.772441640344154 0.8220213747045242 0.0032041583515600935 full 2 WT_1:21/33__IVT_1:830/1075 -0.17694212539156723

I want to know which value is the fold change and the paper said the cutoff is,

Nanocompore: reported p-value<0.01 and GMM log odds ratio>0.5 (for GMM method only).

So, where is GMM log odd ratio, and how to use GMM only, I don't see it on https://nanocompore.rna.rocks/demo/SampComp_usage/

lmulroney commented 1 year ago

Hi JeremyQuo,

The Logit_LOR is the logistical regression log odds ratio value that is used as a secondary threshold with the GMM p-value.

Logan

JeremyQuo commented 1 year ago

Thanks for your answer, I saw your cutoff claimed in Paper is log oddratio> 0.5 Why is not log oddratio> 0.5 orlog oddratio< -0.5

lmulroney commented 1 year ago

Hi @JeremyQuo,

Sorry for the delay before! The description in the Paper may not be as clear as it should be. They used greater than or equal to the absolute value of the log odds ratio as the filtering criteria. You are correct, but taking the abs(LOR) is less error prone to implement in python or R than both conditionals.

Cheers, Logan