Skip to main content
. 2017 Mar 2;45(6):2994–3005. doi: 10.1093/nar/gkx145

Table 1. Properties of selected tools for TAD detection.

Tool Programming language (dependencies) Input Functionality Hierarchical TAD structure Methodology Publication and Link to download
Armatus C++ (C++11, boost) n × n interaction matrix -gzipped Calls TADs on raw and normalized matrices No Consensus of multi-resolution TAD analysis obtained from dynamic programming algorithm (37) https://github.com/kingsfordgroup/armatus/releases
Arrowhead Java (none) hic format (Not from interaction matrix) Calls TADs-part of JuiceBox which has several other functionalities Yes Identification of domain corners using heuristics (14) http://www.aidenlab.org/commandlinetools/
DomainCaller perl, matlab (none) n × (n+3) interaction matrix Calls TADs No Directionality index combined with hidden Markov model (12) http://chromosome.sdsc.edu/mouse/hi-c/download.html
HiCSeg R, C n × n interaction matrix Calls TADs No Maximum likelihood segmentation using dynamic programming algorithm (38) https://cran.r-project.org/web/packages/HiCseg/index.html
TADbit Python (Scipy, numpy, matplotlib, imp, mcl, chimera) n × n interaction matrix Mpas reads, normalizes and plots IF matrices. Calls and plots TADs. TAD clustering. 3D modelling No BIC-penalized breakpoint detection algorithm based on probabilistic interaction frequency model bioRxiv 036764 https://github.com/3DGenomes/tadbit/
TADtree Python (Scipy, numpy) n × n interaction matrix Calls TADs Yes Hierarchical segmentation based on empirical distributions of interaction frequencies within TAD (43) http://compbio.cs.brown.edu/projects/tadtree/
TopDom R n × (n + 3) interaction matrix Calls TADs No Heuristic for breakpoint detection combined with statistical filtering of false-positives (41) http://zhoulab.usc.edu/TopDom/