On reconstructing species trees from gene trees in term of duplications and losses

This paper studies various properties of the least common ancestors mapping, the duplication and mutation costs, and the complexity of finding a species tree from gene trees. 1 Introduction Since DNA sequences have become easier to obtain, emphasis has been placed on constructing gene trees and from these, reconstructing evolutionary trees for species (called species trees) in the evolutionary biology(8, 18, 6). The current strategy for reconstructing species trees is based on the separate consideration of distinct gene families represented by homologous sequences. The homologous sequences are assumed to evolve in the same way as species. However, because of the presence of paralogy, sorting of ancestral polymorphism and horizontal transfer, gene trees and species trees are often inconsistent(19, 23, 26, 28). Hence, a major problem that arises is how to reconcile different, sometimes contradictory, gene trees into a species tree(7). This problem has been studied extensively for...