Hello! #C07URFZF074 I have a question. Do you mind explaining me how removal of redundant families works?

Hello Jose! Currently we do an hmmsearch of all protein family representatives, against all family HMMs. If a match between a family rep and a family hmm (disregarding self hits of course) is above the user set length and e-value thresholds, we keep the family with the most members. This is not perfect, because it might remove superfamilies or smaller families, except if you choose a very strict length threshold (e.g., >0.95). We aim to add an extra local module after hmmsearch in the near future, where we are going to check for Jaccard Index similarities between match families instead. If you are interested in learning more, I will give a bytesize talk about proteinfamilies today at 13:00 CET. It will be available on YouTube afterwards.

Evangelos, Thank you very much for the clear answer. If you could send me the youtube link, it would be great! Thanks. ( I am new in all this NF community) I would suggest to be a little bit more explicit about this addition. For example, I was using the pipeline (which is great btw thanks!) and I found out of this reduction of the families, then I found in the hidden parameter of the pipeline that the removal of families redundancy was set by default. but it took me a while hahaha and then I also read about this in the output doc in the nf-core web page. Regards, Juan

I will post the link as soon as the video is up! Thanks for the nice words as well. It's a lot to digest as a beginner, I understand ;) It takes some time to get accustomed to each pipeline's parametrization, but then it becomes smooth sailing! Even better, you might be able to contribute to your favourite pipelines in the future :D