As I understand, proteins/genes are considered homologous if they share a common ancestor, and sequence based analysis is just a way of trying to define the presence and type of this homologous relationship.
Presumably, many proteins today evolved from a small subset of proteins that were present at the beginning of protein-based biology. Does this mean that most proteins could be considered homologous to each other, unless they evolved from a completely different 'first protein'? I.e. the entire global proteome could be split into N orthologous groups, where N is a relatively small population of 'first proteins'.
I'm just trying to decide whether its correct to say we can use general homology as a predictor of functional relationships (with orthologous relationships being a better predictor than paralogous relationships). If the definition of homology is that loose and the homology groups that large, I'm not sure I can say that, as the functional stringency of given homology group is probably incredibly lax (if, for ?
An alternative statement would be 'proteins that are more homologous are more likely to share function', i.e. if they have HSPs with higher BLAST bitscores for example. But I've also seen that homology is binary and doesn't exist on a scale, so this statement may also be incorrect.