Accurate phylogenetic placement relies on good alignment of both the reference sequences used to build the tree, and the query sequences to be inserted. We can help on the reference tree side of things by making the default alignment algorithm more robust (by using mafft-linsi). Obviously for larger alignments this wont be feasible. But its something to look into to improve reference alignments when we have, say, <500 sequences?
Accurate phylogenetic placement relies on good alignment of both the reference sequences used to build the tree, and the query sequences to be inserted. We can help on the reference tree side of things by making the default alignment algorithm more robust (by using mafft-linsi). Obviously for larger alignments this wont be feasible. But its something to look into to improve reference alignments when we have, say, <500 sequences?
Joel