add benchmark EP interaction paper

README.md

@@ -364,6 +364,7 @@ A paper from Genome Research [Ubiquitous heterogeneity and asymmetry of the chro
 * [TargetFinder](https://github.com/shwhalen/targetfinder). paper: [Enhancer–promoter interactions are encoded by complex genomic signatures on looping chromatin](http://www.nature.com/ng/journal/v48/n5/full/ng.3539.html)
 * [C3D](https://github.com/mlupien/C3D) Cross Cell-type Correlation in DNaseI hypersensitivity. calculates correlations between open regions of chromatin based on DNase I hypersensitivity signals. Regions with high correlations are candidates for 3D interactions. It also performs association tests on each candidate and adjusts p-values. 
 * [ABC](https://www.biorxiv.org/content/10.1101/529990v1) Activity-by-Contact model of enhancer specificity from thousands of CRISPR perturbations. Blog post https://jesseengreitz.wordpress.com/2019/02/10/preprint-activity-by-contact-model/
+* [A curated benchmark of enhancer-gene interactions for evaluating enhancer-target gene prediction methods](https://genomebiology.biomedcentral.com/articles/10.1186/s13059-019-1924-8) "We use BENGI to test several published computational methods for linking enhancers with genes, including signal correlation and the TargetFinder and PEP supervised learning methods. We find that while TargetFinder is the best-performing method, it is only modestly better than a baseline distance method for most benchmark datasets when trained and tested with the same cell type and that TargetFinder often does not outperform the distance method when applied across cell types."
 
 
 ### Allele-specific analysis