Hello, I had a quick question about this pipeline. I've been using scripts stitched together to do something similar (MMseqs -> align -> HMM creation -> HHBlits HMM vs HMM alignment -> MCL clustering) to create "subfamilies" (MMSeqs) within Families (HMM vs HMM alignment MCL clusters). I was wondering does the version of HHBlits incorporate the secondary structure that HHBlits offers? I haven't been able to get that to work in my own version and would like to try this pipeline out.

Hi Zachary! #C07URFZF074 is not using

hhblits

, mainly because I was facing problems when trying to create a module for

hhbuild

(since you need the database with family HMMs first). Instead, it uses a "strict" threshold for

hmmsearch

to find potential duplicate families to remove (searches family representative sequences against all families HMMs), and a more "relaxed" threshold for

hmmsearch

again, to merge families. Do try it out and let me know how well this works for you 😉 We tested this method in various rounds of benchmarking and seemed to be working great! (relevant publication here: https://academic.oup.com/gigascience/advance-article/doi/10.1093/gigascience/giag009/8435307)