How to choose a reference genome

[yjiakang]

Hello, thanks for the wonderfull tool you have developed for exploring MGE! As you mentioned in your manuscript, the choice of reference genome is important (isolates should share at least 98.5% nucleotide identity with the reference genome), so before analyses, I want to consult you how should I choose my reference genome. Here I have downloaded ~200 genomes of one bacterial species, and from the phylogenetic tree I found that they were divided into five clades, so I want to do some analyses about mobile genetic elements of the five clades, and compare them.

I intend to perform MGE analysis of the five clades separately using your MGEfinder, however, I am a little confused about which genome can be used as the reference genome for each clade, can you help me? Thanks in advance. Best, jk yin

[durrantmm]

Hello!

Great question. An ideal reference genome would be a historical isolate that is a common ancestor to your strains of interest. For example, if you are looking for antibiotic resistance mutations, an isolate that predates the widespread use of antibiotics would be ideal.

I would use a tool like FastANI to determine how closely related your 200 isolates are to a selection of possible reference genomes. If they are all above the 98.5% threshold, just use one reference genome for all 200 isolates. If there is more diversity than this, you may want to use a different reference genome for the different clades, but this would make it more difficult to compare the results across the five clades.

I hope that helps! Let me know if you have any more questions.

Matt

[yjiakang]

Thanks for your valuable suggestions, fastANI showed that all my 200 genomes similarity are above 98.5% threshold. jk yin