Re: [Gusdev-gusdev] parseBlastFilesForSimilarity.pl

[Alberto D. <da...@io...>]

Hi Steve,

Yesterday Poliana presented an internal seminar about GUS (focused on
tables used by GBParser and GlimmerM), then some questions arised, maybe
you (or someone else) could help:

Why we need several "string" columns in "DoTs.NASequenceImp":

DoTs.NASequenceImp
string1
string2
string3
string4
string5

What should be stored in the columns of DoTs.FeatureImp :

DoTs.FeatureImp
int1        ...  int6   
tinyint1  ...  tinyint6         
float1     ...  float6  
text1   
string1   ...  string50

Finally, (I think this is for Bindu) does the current version of the
"ImportPlasmoDBPrediction" plugin support more than one ORF predicted by
GlimmerM ? in other words, GlimmerM usually present several rows of
predicted ORFs, each one with a score probability, and there might be
more than one ORF predicted... and more than one can be a valid ORF... 

Thanks, Alberto

On Thu, 2005-02-17 at 15:59 -0500, Steve Fischer wrote:
> luckily we are already adding that in 3.5
> 
> see AnalysisAlgorithm in the 3.5 roadmap: 
> http://www.gusdb.org/wiki/index.php/Gus3.5RoadMap
> 
> steve
> 
> Josef Jurek wrote:
> 
> >Concerning the DoTS.Similarity table which has been discussed
> >recently, what we at the Preuss lab are really hurting for
> >is a way to distinguish different Blast analyses (different
> >blast parameters) of the same subject and query sequences.
> >A table as simple as:
> >
> >   id                             number(10)
> >   name                           varchar2(255)
> >   parameters                     varchar2(255)
> >
> >so that one could store something as simple as:
> >
> >   'Param Set 001'
> >   '-p blastx -M BLOSUM50 -G 13 -E 2 -e .01 -W 3'
> >
> >and a way to link these entries to the rows of
> >DoTS.Similarity, perhaps in the way that
> >DoTS.NASequenceKeyword links rows of DoTS.NASequence
> >and DoTS.Keyword, would work.  If we were to go ahead and
> >add the table to the schema, we might want to actually
> >list the parameters individually with a
> >comment string, perhaps like:
> >
> >
> >   name                          varchar2(255)
> >   ...
> >   expectation_threshold         float(22)
> >   gap_penalty_existence         number(10)
> >   gap_penalty_extension         number(10)
> >   ...
> >   comment                       varchar2(255)
> >
> >Are not other people on the list
> >missing the ability to distinguish
> >blast results by parameters?
> >
> >Thanks, Josef
> >
> >
> >
> >Josef Jurek, Ph.D.
> >
> >Daphne Preuss Laboratory
> >Molecular Genetics and Cell Biology
> >The University of Chicago
> >ju...@cs...
> >
> >voice: (773) 834-3985
> >fax:   (773) 702-6648
> >
> >