
This is the software package mGene, to reproduce the results for gene
finding performance on nematodes presented in [1]. Please, do not
expect this version to run at this point. It needs some modifications
especially in terms of path configuration. If you are intereseted in
reproducing our results, please contact us directly, and we will help
you to get this version to run. It is also not the purpose of this
release to allow the application of mGene to other data. The
documentation of the code is therefore rather sparse.  Please visit
our webservice (http://mgene.org/web) or download the current version
of mGene (which is fully supported) at http://mgene.org/downloads



The scripts run with Matlab and CLPEX.



%%% Category 1: mGene.init

For training the ab-initio version of mGene (mGene.init) we used the matlab script: 
experiments/mGene.init_train.m

For predicting with mGene.init we used the matlab script:
experiments/mGene.init_predict.m



%%% Category 2: mGene.multi

For training the version of mGene that uses multiple genome alignments (mGene.mulit) we used the matlab script: 
experiments/mGene.mulit_train.m

For predicting with mGene.multi we used the matlab script:
experiments/mGene.multi_predict.m



%%% Category 3: mGene.seq

We used EST, cDNA and protein alignments to improve prediction accuracy.
However we did not train a new model, but rather used the mGene.init model.
The additional information was only considered during training. 
For predicting with mGene.seq we used the matlab script 
experiments/mGene.seq_predict.m





**References** 

[1] Schweikert, G., Zien, A., Zeller, G., Behr, J.,
Dieterich, C., Ong, C., Philips, P., Bona, F. D., Hartmann, L.,
Bohlen, A., Kroeger, N., Sonnenburg, S., and Raetsch, G. (2009)
mGene: Accurate computational gene finding with application to
nematode genomes. Genome Research, under review.