Context

We want to evaluate:

• the performance of ABC method [1] to predict Enhancer-Gene relationships, over:
  • the CRISPRi-FlowFISH set
  • the BENGI sets
• the performance of Average Rank method [2] to predict Enhancer-Gene relationships, over:
  • the CRISPRi-FlowFISH set
  • the BENGI sets

In particular, we needed to understand:

• how to use the BENGI sets
• how to use the Average Rank method

This is the subject of this section.

Introduction

These pages gather guidelines to use Moore et al. BENGI benchmarks, mainly for GM12878.

For most of them, the code can be directly adapted to re-run the steps anywhere needed. For some others, one should use the dedicated notebook. All notebooks can be found at scripts/notebooks/ipynb/.

References

[1] Moore, J.E., Pratt, H.E., Purcaro, M.J. et al. A curated benchmark of enhancer-gene interactions for evaluating enhancer-target gene prediction methods. Genome Biol 21, 17 (2020)

[2] Activity-by-Contact model of enhancer-promoter regulation from thousands of CRISPR perturbations. Fulco CP, Nasser J, Jones TR, et al. Nat Genet. 51(12):1664-1669 (2019)