paper | download | supplemental data | tutorial


Improved Gene Tree Error Correction in the Presence of Horizontal Gene Transfer
Mukul S. Bansal*, Yi-Chieh Wu*, Eric J. Alm, and Manolis Kellis.
Bioinformatics. 2015. doi: 10.1093/bioinformatics/btu806

Address correspondence to: Mukul Bansal (mukul at and Manolis Kellis (manoli at

Additionally, if you use the default module for computing the test statistic, please cite
Alexandros Stamatakis. RAxML-VI-HPC: Maximum Likelihood-based Phylogenetic Analyses with Thousands of Taxa and Mixed Models. Bioinformatics 22(21):2688-2690, 2006


TreeFix-DTL is a phylogenetic program for reconstructing highly accurate prokaryotic gene trees. TreeFix-DTL uses the species tree topology to guide the reconstruction of the gene tree and balanaces sequence and species tree information through a statistical hypothesis testing framework. TreeFix-DTL operates under the assumption that multiple gene tree topologies are "statistically equivalent" (as evaluated using a likelihood ratio test), and the optimal topology among this set is one that is most parsimonious (where the cost of the topology is dependent on the inferred gene duplication, horizontal gene transfer, and gene loss events). Specifically, given a (maximum likelihood) gene tree, a multiple sequence alignment, and a rooted species tree, TreeFix-DTL finds an alternative gene tree topology that minimizes the duplication-transfer-loss reconciliation cost while still being supported by the sequence data.

The TreeFix-DTL package includes the Python source code, modified RAxML source code, the ranger-dtl-U executable, as well as several library interfaces for Python. A detailed README and sample dataset are also included.


Likelihood models

By default, TreeFix-DTL computes p-values based on the Shimodaira-Hasegawa (SH) test statistic with RAxML site-wise likelihoods. This is included in the main TreeFix-DTL package.

Reconciliation models

By default, TreeFix-DTL uses ranger-dtl-U reconciliation and computes the duplication-transfer-loss cost (with default costs D=2, T=3, L=1). This is included in the main TreeFix-DTL package.


A fairly thorough tutorial with detailed installation instructions, descriptions of command line options, and step-by-step instructions for using TreeFix-DTL is available.

Supplemental data

In addition to the evaluations discussed in our paper, we evaluated TreeFix-DTL on a dataset of 100 predominantly prokaryotic species sampled broadly from across the tree of life. (This dataset was first analyzed using AnGST.) Additional datasets available upon request.

Last updated 06/22/15.