The phylogenetic relationships between recent Elephantidae (Proboscidea, Mammalia), that is to say extant elephants (Asian and African) and extinct woolly mammoth, have remained unclear to date. The prevailing morphological scheme (mammoth grouped with Asian elephant) is either supported or questioned by the molecular results. Recently, the monophyly of woolly mammoths on mitochondrial grounds has been demonstrated (Thomas, et al., 2000), but it conflicts with previous studies (Barriel et al., 1999; Derenko et al., 1997). Here, we report the partial sequencing of two mitochondrial genes: 128 bp of 12S rDNA and 561 bp of cytochrome b for the Lyakhov mammoth, a 49,000-year-old Siberian individual. We use the most comprehensive sample of mammoth (11 sequences) to determine whether the sequences achieved by former studies were congruent or not. The monophyly of a major subset of mammoths sequences (including ours) is recovered. Such a result is assumed to be a good criterion for ascertaining the origin of ancient DNA. Our sequence is incongruent with that of Yang et al. (1996), though obtained for the same individual. As far as the latter sequence is concerned, a contamination by non-identified exogenous DNA is suspected. The robustness and reliability of the sister group relation between Mammuthus primigenius and Loxodonta africana are examined: down-weighting saturated substitutions has no impact on the topology; analyzing data partitions proves that the support of this clade can be assigned to the most conservative phylogenetic signal; insufficient taxonomic and/or characters sampling contributed to former discordant conclusions. We therefore assume the monophyly of "real mammoth sequences" and the (Mammuthus, Loxodonta) clade.
Copyright 2002 Elsevier Science (USA)