Skip to content

Triangular Alignment (TAME): A Tensor-based Approach for Higher-order Network Alignment

Notifications You must be signed in to change notification settings

shmohammadi86/TAME

Repository files navigation

Triangular Alignment (TAME) is a new formulation for the network alignment problem that uses network motifs as the driving factors to align a pair of given graphs. In order to install TAME, simply download the latest version and run:

make

To run TAME, the bare minimum configuraion is as follows:

tri-match -t <type> -G <1st_net> -H <2nd_net>

where the network files are in tab separated format by default. In this sense, each file contains lines in the form: "src_node dst_node." Additional parameters can be provided to config different aspects of TAME. To get a comprehensive list of parameters, type:

./tri-match --help

There are three example scripts to help you get a sense of how to start using TAME:

  1. sparse_run.sh: aligns yeast vs human interactome with sequence similarity in sparse mode.
  2. dense_run.sh: aligns yeast vs human interactome with sequence similarity in dense mode.
  3. dense_run_noPrior.sh: aligns yeast vs human interactome without sequence similarity in dense mode.

Cite

Mohammadi, S., Gleich, D. F., Kolda, T. G., & Grama, A. (2016). Triangular Alignment (TAME): A Tensor-based Approach for Higher-order Network Alignment. IEEE/ACM Transactions on Computational Biology and Bioinformatics (http://doi.org/10.1109/TCBB.2016.2595583)

About

Triangular Alignment (TAME): A Tensor-based Approach for Higher-order Network Alignment

Resources

Stars

Watchers

Forks

Packages

No packages published