Skip to content. | Skip to navigation

Personal tools

Sections
You are here: Home / Teams / Physical Biology of Chromatin - D. Jost / Publications / 3DGenBench: a web-server to benchmark computational models for 3D Genomics.

3DGenBench: a web-server to benchmark computational models for 3D Genomics.

Polina Belokopytova, Emil Viesná, Mateusz Chiliński, Yifeng Qi, Hossein Salari, Marco Di Stefano, Andrea Esposito, Antonio Conte, Andrea M Chiariello, Vladimir B Teif, Dariusz Plewczynski, Bin Zhang, Daniel Jost, and Veniamin Fishman (2022)

Nucleic Acids Res.

Modeling 3D genome organisation has been booming in the last years thanks to theavailability of experimental datasets of genomic contacts. However, the field iscurrently missing the standardisation of methods and metrics to comparepredictions and experiments. We present 3DGenBench, a web server available athttps://inc-cost.eu/benchmarking/, that allows benchmarking computational modelsof 3D Genomics. The benchmark is performed using a manually curated dataset of 39capture Hi-C profiles in wild type and genome-edited mouse cells, and fivegenome-wide Hi-C profiles in human, mouse, and Drosophila cells. 3DGenBenchperforms two kinds of analysis, each supplied with a specific scoring module thatcompares predictions of a computational method to experimental data using severalmetrics. With 3DGenBench, the user obtains model performance scores, allowing anunbiased comparison with other models. 3DGenBench aims to become a reference webserver to test new 3D genomics models and is conceived as an evolving platformwhere new types of analysis will be implemented in the future.

 
automatic medline import

Document Actions