Max-Planck-Institut für Informatik
max planck institut
mpii logo Minerva of the Max Planck Society

Mtreemix - Mutagenetic trees mixture models

Current version

1.3.00 Mai 2007

What's new?


Mtreemix is a software package to estimate mixture models of mutagenetic trees from observed cross-sectional data. Mutagenetic tree mixtures are probabilistic models that have been designed to describe evolutionary processes that are characterized by the accumulation of genetic changes. Mtreemix has been applied to model the development of drug resistance-associated mutations in the HIV genome and the accumulation of chromosomal gains and losses in tumor development. (For more details, see the references below.)

The Mtreemix C code has been developed on the basis of LEDA and uses CLAPACK. You can download the Mtreemix source code and precompiled Solaris and Linux binaries under the licence terms stated below. The precompiled binaries can be used directly without any further requirements. (But for recompiling the source code the LEDA and CLAPACK packages are necessary.)



mtreemix_fit       - estimate K-mutagenetic trees mixture model from data
mtreemix_fit0      - estimate null model (independence of events) from data
mtreemix_fit1      - estimate single mutagenetic tree from data
mtreemix_distr     - compute entire distribution encoded by model
mtreemix_time      - estimate pattern waiting times
mtreemix_sim       - draw patterns from model
mtreemix_wait      - draw patterns and their waiting times from model
mtreemix_loglike   - log-likelihood computation
mtreemix_transprob - transition probabilities between all patterns
mtreemix_select	   - model selection (cross-validation, BIC, modified BIC)
mtreemix_dim       - compute the dimension (the effective number of parameters)
                     of the mixture model
mtreemix_compare   - compare density estimates
mtreemix_bootstrap - analyze tree stability
mtreemix_random    - generate a random mixture model
mtreemix_print     - print model to DOT file
mtreemix_fscore    - Fisher score computation


The mtreemix package comprises

In order to download mtreemix you need to register as a user. Download is subject to your agreement to the Mtreemix End User Licence Agreement.

Download Mtreemix here


When using Mtreemix, please cite:

Further References

  1. N. Beerenwinkel, J. Rahnenführer, M. Däumer, D. Hoffmann, R. Kaiser, J. Selbig, and T. Lengauer. Learning multiple evolutionary pathways from cross-sectional data. J. Comput. Biol. 12(6):584-598, 2005. [RECOMB 2004] [PubMed]

  2. J. Yin, N. Beerenwinkel, J. Rahnenführer, T. Lengauer. Model selection for mixtures of mutagenetic trees. Stat. Appl. Genet. Mol. Biol., Vol. 5, Iss. 1, Article 17. [bepress]

  3. R. Desper, F. Jiang, O.-P. Kallioniemi, H. Moch, C. H. Papadimitriou, A. A. Schäffer. Inferring tree models for oncogenesis from comparative genome hybridization data. J. Comp. Biol. 6(1):37-51, 1999. PubMed

  4. N. Beerenwinkel, M. Däumer, T. Sing, J. Rahnenführer, T. Lengauer, J. Selbig, D. Hoffmann, R. Kaiser. Estimating HIV evolutionary pathways and the genetic barrier to drug resistance. J. Infect. Dis. 191(11):1953-1960, 2005. PubMed

  5. J. Rahnenführer, N. Beerenwinkel, W.A. Schulz, C. Hartmann, A. von Deimling, B. Wullich, and T. Lengauer. Estimating cancer survival and clinical outcome based on genetic tumor progression scores. Bioinformatics 21:2438-2446, 2005. [PubMed]

  6. Tobias Sing, Niko Beerenwinkel. Mutagenetic tree Fisher kernel improves prediction of HIV drug resistance from viral genotype. Bioinformatics 21:2438-2446, 2005. NIPS 2006 - Twentieth Annual Conference on Neural Information Processing Systems. [NIPS Proceedings] [PubMed]

Authors and Contributors


Niko Beerenwinkel
Harvard University
Program for Evolutionary Dynamics
1 Brattle Square
Cambridge, MA 02138