Parallel T-Coffee (PTC) is the first parallel implementation of the TCoffee multiple sequence alignment tool. It has been developed to overcome main limitations of the original method. It is based on the MPI and RMA mechanisms, and it can be run on distributed memory clusters. PTC supports a majority of options provided by TCoffee 3.79, including the 3D-Coffee mode. It can be used to align data sets consisting of hundreds of proteins in reasonable time limits. Parallel T-Coffee is (or has been) developed by Jaroslaw Zola, Xiao Yang, Adrian Rospondek and Srinivas Aluru.

↑Top

• GNU make – to build Parallel T-Coffee we use Make build system. We tested GNU make, there is a chance that standard UNIX-like make will work.
• C++ compiler – we suggest to use the Intel C++ compiler. The GCC set of compilers has been tested as well. In general we expect any compiler that supports C++ 2003 and C99 standards to work.
• MPI library – we suggest to use MPICH2 implementation for its speed and portability. We expect any MPI implementation that supports MPI-2 standard to work. Please note, that typical MPI implementation provides its own wrapper (e.g. mpicxx) over C++ compiler - you should consider this wrapper whenever we refer to C++ compiler. If you want to use RMA mechanism to distribute TCoffee storage you have to be sure that your MPI implementation supports one-sided communication.
• Cluster of workstations – although Parallel T-Coffee can be executed on uni-processor machines we believe that it is better to use a cluster consisting of more than one node with a local storage.

In addition to run 3D-coffee mode

• SAP – structure-structure protein alignment tool, distribution provided by C. Notredame. We suggest to compile it for the particular platform you want to use (this is especially important for 64bit architectures).
• Perl – no suggestions here, just make sure that it is relatively fresh version
• Wget – this tool is provided by majority of Linux distributions and can be installed on many UNIX systems. Just ask your system administrator.
• Internet access over the FTP protocol – in many cases master node will have to connect to the RCSB server using FTP protocol. Make sure that firewalls securing your cluster do allow that kind of activity.

↑Top

The latest version of Parallel T-Coffee is 1.914 released on 29 Aug 2012. Parallel T-Coffee is distributed under GNU General Public License version 2, mpiext component is distributed under Boost Software License version 1.0.

Click here to download

The bundle provides a comprehensive documentation explaining how to install and use PTC. If you find this software useful and you would like to share your experience with us, or you would like to see some additional features in it, please send a feedback to Jaroslaw Zola.

↑Top

Three example data sets aligned by PTC. Please note that the sequential program was not able to align them on Xeon 3GHz with 2GB of RAM.

• PF00074 n=349, Is=0.39, Im=0.06
  • Sequences
  • Alignment
  • Tree
• PF00231 n=554, Is=0.41, Im=0.03
  • Sequences
  • Alignment
  • Tree
• PF00500 n=1048, Is=0.45, Im=0.07
  • Sequences
  • Alignment
  • Tree

↑Top

When using Parallel T-Coffee please cite:

J. Zola, X. Yang, S. Rospondek, S. Aluru: Parallel T-Coffee: A Parallel Multiple Sequence Aligner, In Proc. of ISCA PDCS-2007, pp. 248-253, 2007.

↑Top