Skip to main content

Table 2 The Leiden algorithm outperforms other clustering methods including label propagation, Markov clustering, Infomap, and Louvain for Hi-C data analysis based on the contacts normalized by HiCzin

From: HiCBin: binning metagenomic contigs and recovering metagenome-assembled genomes using Hi-C contact maps

 

No. of groups

No. of contigs

F-score

ARI

NMI

Label propagation

10

5026

0.680

0.599

0.739

Markov clustering

37

4699

0.546

0.471

0.686

Infomap

65

3582

0.772

0.731

0.770

Louvain

14

5241

0.838

0.815

0.842

Leiden (HiCBin)

14

5217

0.908

0.894

0.895

  1. Note: The optimal values of the results are in bold. No. of groups represents the number of valid genome bins. No. of contigs is the number of contigs in the valid bins