Useful websites: https://www.cathdb.info/ http://sfam.cathdb.info/

Part I: Steps followed for naming a superfamily - Look through representative domains as: ‘domain only’ to understand common secondary structures; as ‘domain in chain’ to observe the location of the domain in the chain; as ‘domain in PDB’ to understand the domain’s function and location in the protein. - Check through FunFams/SwissProt/Keywords and refer to the most abundant name when naming. - Check through enzymes (EC number if available), GO terms and species to get a rough idea of domain function. - Refer to Pfam and InterPro entries for general idea of protein domain function and/or structure. - Check through papers associated with PDB entry for better understanding of protein and protein domain structure and/or function . - In ‘Description section’, provide an overview of structure and function. In larger superfamilies, you may have to refer to specific PDB IDs. - Check references are correct: [InterPro:] [Pfam:] [PMID:] [DOI:] - Check other names in the database, either to avoid duplicate names or to identify potential cross-hits - Check names of other domains in the same chain to keep the name similar.

Part II: General observations and tips

Dos • Check other names in CATH to not make duplicates (i.e. make sure the assigned name is unique) • Make superfamily names consistent with other domains of same protein • Start with smaller families until you get the hang of it • For larger superfamily- it is a good idea to check FunFam • When looking at a protein on InterPro, see if there are other domains that don’t have a name yet on the same protein - it will be easy to name that one • Work in groups for larger superfamilies • Choose superfamily entries with FunFams, Pfams, or InterPro associated

Don’ts • Make description without sourcing references • Make description without actually really understanding it • Spend 3 hours on a very small superfamily • Look at every single PDB for big superfamilies • For smaller representative domains, don’t put too much confidence in InterPro/Pfam - it may be better to look at PDB paper for the specific domain • Assume it is the exact same domain if it has good mapping to Pfam • Choose a superfamily entry with no annotation or too many annotation