Currently variants are clustered based on a fixed size cutoff, specified by the --maxdist argument. The fixed cutoff can cause various filtering and reporting issues, as described in #2 (comment). Implementing a clustering method with an adaptive size cutoff could improve these issues.