Title: Rec-I-DCM3: A Fast Algorithmic Technique for Reconstructing Large 
Phylogenetic Trees
Authors: Usman Roshan, Bernard Moret, Tiffani Williams, Tandy Warnow

Estimations of phylogenetic trees are most commonly obtained through the
use of heuristics for maximum parsimony (MP), an NP-hard problem. Although
apparently good heuristics have been devised, even they fail to produce
good solutions in reasonable time for large datasets---today's limit is
unknown, but is probably less than one thousand sequences. In this paper,
we present a promising new divide-and-conquer technique,
Recursive-Iterative-DCM3 (Rec-I-DCM3). This new method belongs to our
family of Disk-Covering Methods (DCMs); it operates by iteratively
dividing the input set of sequences into smaller overlapping subproblems,
solving them using some base method (e.g., neighbor-joining, heuristic MP,
heuristic ML, etc.), and then merging these subtrees into a single,
phylogenetic tree.  Thus, Rec-I-DCM3 is designed to boost the performance
of its base method.  Our new method is composed of a new DCM, which we
call DCM3, but utilizes recursion and iteration as well; the result is a
booster of phylogenetic reconstruction methods that can produce dramatic
improvements for standard heuristics, as well as substantial improvements
over the very best methods (which are harder to improve).  We demonstrate
the power of this new DCM on ten large biological datasets ranging from
1,322 to 13,921 sequences.