Search or download the Clustered sets of related sequences from UniProt.
Highlights:
UniRef combines closely related sequences into a single record to speed sequence searches.
The databases provide complete coverage of sequence space while hiding redundant sequences from view.
The non-redundancy allows faster sequence similarity searches by using UniRef90 and UniRef50 (which merge all records from all source organisms with mutual sequence identity of >90% or >50%, respectively, into a single record).