The fragment classification package (FCP) provides the following classifiers for assigning a taxonomic attribution to metagenomic fragments:
- Naive Bayes (NB): a composition-based, rank-specific classifier
- BLASTN: a homology-based, rank-specific classifier
- NB-BL: a hybrid, rank-specific classifier combining NB and BLASTN
- Epsilon-NB: a composition-based, rank-flexible classifier
- Lowest common ancestor (LCA): a homology-based, rank-flexible classifier
- LCA + epsilon-NB: a hybrid, rank-flexible classifier combining LCA and epsilon-NB
This software is released under the GNU General Public License v3.0.
- Join our mailing list to keep informed about the FCP.
Obtaining the Software
The FCP is available as a gzipped tarball for OSX and Linux users, and a zip file for Microsoft Windows users:
After downloading, uncompress the file and follow the installation instruction in the README file. Under OSX or Linux, the zip file can be uncompressed by typing unzip FCP_1_0_4.zip at the command prompt. Under Windows 7, the zip file can be uncompressed by right-clicking on the file and selecting 'Extract all...' from the popup menu. The FCP has been tested under OS X Snow Leopard, Linux, and Windows 7. To install the FCP under OS X, gcc must be installed which is available as part of the Mac Developer Tools.
Citing the FCP
If you find this software helpful in your research, please cite:
- Parks, D.H., MacDonald, N.J., and Beiko, R.G. (2011). Classifying short genomic fragments from novel lineages using composition and homology. BMC Bioinformatics, 12:328. (Abstract)
The FCP is in active development and we are interested in discussing all potential applications of this software. We encourage you to send us suggestions for new features. Suggestions, comments, and bug reports can be sent to Rob Beiko (beiko [at] cs.dal.ca). If reporting a bug, please provide as much information as possible and a simplified version of the data set which causes the bug. This will allow us to quickly resolve the issue.
v1.0.4 (April 13, 2013)
- removed need for taxonomy file from nb-train
- improved error reporting for nb-train and nb-classify
- add example illustrating use of the naive Bayes classifier independent of the FCP framework
v1.0.3 (July 13, 2011)
- changed NB_install.py to FCP_install.py
- added -protein flag to FCP_install.py to all FCP to be used with RITA
v1.0.2 (June 3, 2011)
- improved robustness of NB_install.py script
- TaxonomicSummary.py now reports both raw and normalized percentages
- removed need for temporary directories when running nb-classify
- fixed major bug in LCA+Epsilon-NB.py which was causing the epsilon parameter to be handled incorrectly
- fixed minor bug in LCA.py where unclassified fragments where reported as 'u' instead of 'unclassified'
v1.0.1 (Apr. 28, 2011)
- revision to NB_install.py to account for genomes without NCBI taxonomy information
v1.0.0 (Jan. 4, 2011)
- initial software release.
The development of this software has been supported by several organizations: