Skip to content

Genome-wide identification of E/G interactions

Results

Genome-wide identification of E/G interactions

Guidebooks
Guidebooks
- Introduction
- Compute summary stats of a bedpe
- Basics
  Basics
- Misc
  Misc
  - Compute correlations
Haemochromatosis
Haemochromatosis
- Introduction
- Data format
- Custom script GO
- Main results
- CHiC data
  CHiC data
- ABC data
  ABC data
BENGI
BENGI
- Introduction
- Notebooks
  Notebooks
  - Convert BENGI datasets to custom bedpe
- Summary statistics
  Summary statistics
  - Summary statistics on BENGI GM12878 benchmarks
  - (R code) Compute summary statistics over BENGI
- Evaluation of a few methods over BENGI GM12878
  Evaluation of a few methods over BENGI GM12878
  - Distance method
    Distance method
    
    Full code
    
    Results
  - DNase-expression correlation method
    DNase-expression correlation method
    
    Full code
    
    Results
  - Average rank method
    Average rank method
    
    Full code
    
    Results Results
    Table of contents
    
    Prerequisites
    
    Not correcting the error in Run-Average-Rank.sh gives same results as in the paper
    
    Correct results
- Evaluation of a few methods over Fulco et al 2019 CRISPRi-FlowFISH
  Evaluation of a few methods over Fulco et al 2019 CRISPRi-FlowFISH
  - Distance method
    Distance method
    
    Full code
    
    Results
  - DNase-expression correlation method
    DNase-expression correlation method
    
    Full code
    
    Results
  - Average rank method
    Average rank method
    
    Full code
    
    Results
ABC
ABC
- Introduction
- ABC over BENGI for GM12878, main results
- CRISPRi-FlowFISH
  CRISPRi-FlowFISH
  - Original ABC predictions over K562
  - Summary ABC over CRiFF
- Notebooks
  Notebooks
- Full codes
  Full codes
  - Test of ABC pipeline over chr22
  - ABC over GM12878 with same accessions as for BENGI
- Meld of results over chr22 vs authors' results
About
About
- Getting started
- About me

Average rank method over BENGI for GM12878¶

Prerequisites¶

Detailed procedure is explained here.

Not correcting the error in `Run-Average-Rank.sh` gives same results as in the paper¶

If we keep the small mistake found in Run-Average-Rank.sh, we find (hopefully!) the very same Precision - Recall curves as authors:

Image: Precision-Recall curves and AUPR with small mistake in Run-Average-Rank.sh

Correct results¶

If we correct the small error found in Run-Average-Rank.sh, we obtain the following results.

Image: Precision-Recall curves and AUPR

Image: Precision-Recall curves