Skip to main content
. 2023 Aug 26;39(9):btad529. doi: 10.1093/bioinformatics/btad529

Table 1.

Comparison of ETNA with existing unsupervised pairwise network alignment methods.

Method Interpretable embeddings Cross-species anchors Combining sequence and topology Directionality Algorithm Code availability
ETNA Yes Orthologs Nonlinearly Bidirectional Autoencoders Python
MUNK (Fan et al. 2019) Yes Orthologs Nonlinearly Directional SVD Python
IsoRank (Singh et al. 2008) No BLAST Linearly Bidirectional PageRank Linux executable
HubAlign (Hashemifar and Xu 2014) No BLAST Linearly Bidirectionala Minimum-degree heuristic algorithm C++
PrimAlign (Kalecky and Cho 2018) No BLAST Linearly Bidirectional Markov chain + PageRank Not available
SANA (Mamano and Hayes 2017) No BLAST Linearly Bidirectionala Simulated annealing Webserver, C++
L-GRAAL (Malod-Dognin and Pržulj 2015) No BLAST Linearly Bidirectional Lagrangian graphlet C++, not working
GHOST (Patro and Kingsford 2012) No GO Does not use sequence Bidirectionala Seed-and-extend with local search C++
NETAL (Neyshabur et al. 2013) No GO Does not use sequence Bidirectionala Greedy search on similarity matrix Webserver
MAGNA++ (Vijayan et al. 2015) No n/a Does not use sequence Bidirectional Genetic algorithm Executable, C++
a

The method has a bidirectional objective function, but the returned output is directional.