Accession Numbers for Nuclear and Mitochondrial Genes

Supplementary Information

Maximum Likelihood Tree for the 8655 bp Data Set

Bootstrap support values based on analyses with the 8655 bp data set. The outgroup was a chimaeric diprotodontian marsupial. The 26 placental taxa from the BRCA1 data set that were concatenated with the 5708 bp data set were as follows: sloth, dugong, African elephant, rock hyrax, aardvark, long-eared elephant shrew, mouse, Old World porcupine, jackrabbit, human, galago, flying lemur, tree shrew, flying fox, round-eared bat, horse, rhino, pig, cow, hippo, humpback whale, pangolin, cat, dog, mole, and hedgehog. The transition to transversion ratio used in maximum likelihood was 1.92. Levels of bootstrap support for Euarchonta + Glires were variable with different methods (e.g., 1% support with minimum evolution; 88% support with maximum likelihood). This variation resulted from differences in the position of the root. However, all methods provided robust bootstrap support (97 to 100%) for Euarchonta + Glires on unrooted trees. Also on unrooted trees, a branch separating Xenarthra + Afrotheria from Laurasiatheria + Euarchonta + Glires was supported at the 100% bootstrap level with all methods.

Results of Monte Carlo Simulations and Parsimony Trees for BRCA1

We assumed the tree in Figure 1b, as well as the shortest parsimony trees for the BRCA1 data set (see below), for purposes of Monte Carlo simulations to determine whether or not we should expect to see systematic rooting errors with parsimony and/or likelihood. Monte Carlo simulations, also known as parametric bootstrapping, use numerical simulation rather than resampling (nonparametric bootstrapping) to generate pseudoreplicate samples¹. Monte Carlo simulations were performed using the program Seq-Gen². Branch lengths for all trees were estimated using maximum likelihood following Hillis et al.¹ and Huelsenbeck³. We used the HKY85 model of sequence evolution,with parameterizations (i.e., base frequencies, transition to transversion ratio) derived from the actual BRCA1 data set. 100 data sets were simulated for each tree. Data sets were the same size as the original (2947 nucleotides). Seq-Gen settings corresponding to the above were as follows: -mhky –t2.22 –n100 –l2947 –f0.365,0.188,0.218,0.229.

1. 
   Hillis, D. M., Mable, B. K. & Moritz, C. in Molecular Systematics (eds Hillis, D. M., Moritz, C. & Mable, B. K.) 407-492 (Sinauer, Massachusetts, 1996).
   
2. 
   Rambaut, A. & Grassly, N. C. Seq-Gen: An application for the Monte Carlo simulation of DNA sequence evolution along phylogenetic trees. Comput. Appl. Biosci. 13, 303-306 (1997).
   
3. 
   Huelsenbeck, J. P. Is the Felsenstein Zone a fly trap? Syst. Biol. 46, 69-74 (1997).
   

1. 
   Rambaut, A. & Bromham, L. Estimating divergence dates from molecular sequences. Mol. Biol. Evol. 15, 442-448 (1998).
   

2. 
   Gheerbrant, E., Sudre, J. & Cappetta, H. A Palaeocene proboscidean from Morocco. Nature 383, 68-70 (1996).
   

3. 
   McKenna, M. C. & Bell, S. K. Classification of Mammals Above the Species Level (Columbia Univ. Press, New York, 1997).
   

4. 
   Amrine, H. M. & Springer, M. S. Maximum likelihood analysis of the tethythere hypothesis based on a multigene data set and a comparison of different models of sequence evolution. J. Mammal. Evol. 6, 161-176 (1999).
   

5. 
   Waddell, P. J., Cao, Y., Hasegawa, M. & Mindell, D. P. Assessing the Cretaceous superordinal divergence times within birds and placental mammals by using whole mitochondrial protein sequences and an extended statistical framework. Syst. Biol. 48, 119-137 (1999).
   

6. 
   Swofford, D. L. PAUP*. Phylogenetic Analysis Using Parsimony (* and Other Methods). Version 4, (Sinauer, Sunderland, MA, 1998).