composerV
is a suffix-tree based implementation of a variety of alignment-free sequence comparison methods. Using the ndexing power of the suffix tree the package makes it easy to implement and add scores based on word composition of suffix trees.
- (Nov. 14, 2009) build 1.0.9: There is a memory-leak problem when the `project' module is used, so I added some python scripts to call separately the client executable. This seems to work without any problems, and the command line interface is nicer. Of course, there is no distinction on the running time since all calculations are done in the `client' module. I did do some work on memory management and logging, so at least now there is a way to track memory allocation. I am seriously thinking of removing the C implemented `project' module and distribute the program as a python extension with a couple of easy-to-use python scripts. We'll see ...
- (Nov. 4, 2009) build-1.0 The first issue of the program! It comes with a suffix tree module, an I/O module and a distance matrix module. All wrapped in two executables: client and project. The former computes the distance between two sequences and the latter can be called in a directory of fasta files to compute a distance matrix. Have a look at the README and the documentation. Here is a list of experiments (using phylip's neighbor and drawtree in sequence, and using the splitStree program).
CV-Tree
An efficient implementation of the CV-Tree method is available for use here.
Verbumculus
Verbumculus is a program for identifying and annotating all maximal subwords of a give texts string. This problem arises frequently in sequence analysis in molecular biology, in pattern discovery and data compression. A description of the tool and a web-based interface can be found here. The source code of a C-implementation combined with SQLite for result analysis is available for download here.
home