Results


ClassyFlu, BLAST and a phylogenetic approach have been applied to full gene sequences or fragments of the HA gene, from the NCBI Influenza Virus Resource and the GISAID EpiFlu influenza database with a release date between Jan. 2011 and Sep. 2012. This is a set of 1,033 sequences, 179 of them from GISAID. The results are provided in the following table: ClassyFlu vs. BLAST vs. PHYLO
These results are described in the article.


Algorithm


The algorithm was developed as part of the diploma thesis of Sandra Van der Auwera:
ClassyFlu-algorithm.pdf [3 MB]


Source code


The following archive contains all source code files to run the classification with hmmClassify.pl locally.

ClassyFlu.zip [1 MB]

The following archive contains all source code files to perform the HMM profile training with hmmTraining.pl locally.
This training algorithm was used for the iterative, discriminative adjustment of the HMM parameters.

hmmTraining.zip [16 KB]

Additional example and README files are added to understand and run the program.


Training data


Accession numbers of NA training genes: allNtrain.lst