From Bioinformatics Software
Revision as of 21:18, 11 July 2015 by Dparks (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search

The fragment classification package (FCP) provides the following classifiers for assigning a taxonomic attribution to metagenomic fragments or assembled scaffolds:

  • 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.

Obtaining the Software

After downloading, uncompress the file and follow the installation instruction in the README.md file. Under OSX or Linux, the zip file can be uncompressed by typing unzip FCP_1_0_7.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)

Contact Information

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.

Version History

v1.0.7 (July 12, 2015)

  • fixed AddModel.py should it accepts the desired n-mer size of new models

v1.0.6 (August 12, 2014)

  • fixed bug with NB-BL.py which caused it to have an output format incompatible with the rest of the FCP suite

v1.0.5 (May 30, 2014)

  • changed default k-mer size from 10 to 8 in response to the rapidly expanding number of reference genomes
  • minor code improvements

v1.0.4 (April 13, 2013)

  • removed need for taxonomy file from nb-train
  • improved error reporting for nb-train and nb-classify
  • added 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: