sanger-pathogens/Roary

MCL clustering

mweberr opened this issue · 0 comments

Hi,
I have an issue understanding how Roary clusters the sequences after CDHIT step.
The paper states : " Sequences are then clustered with MCL ([Enright et al., 2002]), and finally, the pre-clustering results from CD-HIT are merged together with the results of MCL."

I have not worked with MCL yet. Does MCL cluster the blast all-against-all network or the raw input sequences ?
Thank you for clarification.

Best, Michael