Genome Dereplication

Dereplicate a set of genomes using a threshold for ANI (Average Nucleotide Identity) or AAI (Average Amino Acid Identity).

Use this application to identify representatives from a collection of (redundant) genomes


A collection of genomes (assemblies), typically 10 or more, where some are suspected to belong to the same genomic group.


  • Index method: Select the search engines for the all-vs-all comparisons. We strongly recommend using the default, Fast, which uses Diamond-based AAI and FastANI. If you would like your results to be directly comparable to BLAST-based AAI and ANI, you should select Sensitive.

  • Distance metric: Select the type of distance metric to use. If you dereplicate your genomes at or below the species level, use Average Nucleotide Identity or ANI. If you want to dereplicate your genomes at or above the genus level, use Average Amino Acid Identity or AAI.

  • Metric threshold: Minimum percentage value of AAI or ANI to group genomes together. For reference, species-level groups typically correspond to ANI ≥ 95%, whereas genus-level groups typically correspond to AAI ≥ 60%. You can increase this value to select smaller, more redundant groups; or increase it to generate larger groups.

  • Representative Criterion: Criterion to pick cluster representatives. Once MiGA identifies non-redundant groups of genomes, it will also propose a single genome from each group to serve as the representative of the group. If your genomes vary widely in quality, it's a good idea to pick Highest genome quality. However, if all genomes in the collection are of comparable quality (e.g., all are complete genomes), you can ask MiGA to select instead the genome closest to the center of the cluster, Cluster medoids.