Domains, Folds, Families, Superfamilies

[ennui]

Bioinformatics is giving me a headache! Lots of authors mention the terms 'protein fold', 'protein superfold', 'protein family', 'protein superfamily' and 'protein domain' interchangeably.

 

How do you guys use these terms? What's standard in the protein community?

[CharonY]

Actually each refers to something different, so there is no way that they are used interchangeably. However, what is used depends strongly on context. A protein fold, for instance, is precisely that, one of the many possible folds found in proteins. Superfolds, on the other hand, refer to one of the roughly nine or basic folds that have no sequence or function relatedness but which dominate structural databases.

 

Families on the other hand usually refer to relatedness rather than the structure itself, though protein families can be built either on sequence or structural similiartities. While often an phylogenic relationship is inferred, it may be solely based on other aspects as well. In other words, there are different type of protein family classification methods.

 

A protein domain is usually defined as a stretch of protein with biological interest, as e.g. active center, co-factor binding area, etc. It may contain one or (usually) more folds.

 

Again, each refers to another aspect, depending on context. They cannot be used interchangeably, but they may describe the same structure, only from a different viewpoint. E.g. a given protein domain, may be used as a classified to define a protein family. I would try to read the papers with that in mind once more carefully.

[ennui]

Thank you! God, I've had no end of trouble with protein classification. My teacher has been away and journals haven't been a help. That's cleared a lot of stuff up. Gracias!!