search

broadinstitute / ABC-Enhancer-Gene-Prediction

Cell type specific enhancer-gene predictions using ABC model (Fulco, Nasser et al, Nature Genetics 2019)
MIT License
192 stars 56 forks source link

what's the preferred setting when mean associated enhancers is lower than 2 ? #29

Open lmangnier opened 3 years ago

lmangnier commented 3 years ago

Hi there,

I ran your code on my own Hi-C data for 3 cell lines (Neural progenitors cell lines, Derived Neurons and Astrocyte glials) using the methodology you describe, but I find a average number of associated distal enhancers per genes lower than 2 (about 1.80 for my 3 cell lines). You mentioned in one of the last section of your github, some adjustments can be done, quantile normalization or modify the ABC-Score threshold. I used default settings of you code, KR normalization on 10 Kb resolution, however when I VC normalize my data I obtain a average number of distal enhancer in the required range of values. I would like to know what is the preferred setting to adjust? Do quantile normalization, decrease the threshold, use a particular normalization ?

Thanks,

Loic.

thouis commented 3 years ago

@jnasser3 - do you have an answer for this?