Usually up to 3 rounds of clustering, depending on the size of input and classes, would be enough to identify the homologs. For large datasets with thousands of sequences, further iterations of clustering can be helpful. For such cases, the sub-workflow concept that is recently provided by the Galaxy team is handy. For enhancing the invocation of further rounds, we have encapsulated initial and iterative rounds of GraphClust2 as sub-workflow. This encapsulates the collection of individual tools and provides it as a single sub-workflow. By connecting the initial fast clustering with custom numbers of iterative sub-workflows a custom number of iterations would be achieved.