[Bioperl-l] Bio::Tools::MZEF

Hilmar Lapp hlapp@gmx.net
Thu, 03 Aug 2000 14:47:15 +0200

There is now also a Bio::Tools::MZEF for parsing MZEF prediction results into
GeneStructure objects. It works pretty much like the Genscan module, apart
from the inherent limitation that MZEF would not predict individual genes and
hence you basically get one single container with all exons MZEF predicted.
These are Bio::Tools::Prediction::Exon instances, like for the Genscan module,
so you have all the scores at hand, as well as the likeliest frame.

In addition it lets you specify -strand => -1 at initialization, to indicate
that you reversed the sequence prior to feeding it to MZEF. This will then be
reflected in the strand() attribute of the exon objects, and start/end
coordinates will be translated to the plus strand. There seems to have been a
version that printed an indication of the strand to the output, but the
version I have doesn't (and it isn't interactive), and it seems to be the
newer one.

The $gene->cds() method here doesn't obviously work as well, even not with
TRUE as argument for correcting the phase, because there are no initial exons,
and exons from several genes may be put together, depending on the length of
the input sequence. This needs more testing. It could be that if you provided
a sequence that covers only one gene $gene->cds(1) may still work. If there
were several genes, it almost certainly fails to report a sequence free of
intervening terminators. It could also be that one will need to correct every
exon on its own, taking into account the frame() attribute.

Hilmar Lapp                                      email: hlapp@gmx.net
NFI Vienna, IFD/Bioinformatics                   phone: +43 1 86634 631
A-1235 Vienna                                      fax: +43 1 86634 727