[Bioperl-l] Re: SiteMatrix

[Stefan Kirov]

Got it!
It is actually a design issue in SiteMatrix- the constructor should 
rather call _initialize instead of doing  everything by itself. Right 
now since Psm is SiteMatrix, SiteMatrix is trying to create the 4 
vectors (A,G,C,T) even though there is no such need (in the case of 
parsing mast for example). I will apply a quick fix to check for the 
input more properly and later I will split the constructor into two 
subroutines, which should be a permanent solution.
Thanks for pointing the problem Heikki.
Stefan

Heikki Lehvaslaiho wrote:

>These are the error messages:
>
>t/psm........................ok 39/42Use of uninitialized value in split
>at /home/heikki/src/bioperl/core/blib/lib/Bio/Matrix/PSM/SiteMatrix.pm
>line 177, <GEN2> line 172.
>Use of uninitialized value in split at
>/home/heikki/src/bioperl/core/blib/lib/Bio/Matrix/PSM/SiteMatrix.pm line
>178, <GEN2> line 172.
>Use of uninitialized value in split at
>/home/heikki/src/bioperl/core/blib/lib/Bio/Matrix/PSM/SiteMatrix.pm line
>179, <GEN2> line 172.
>Use of uninitialized value in split at
>/home/heikki/src/bioperl/core/blib/lib/Bio/Matrix/PSM/SiteMatrix.pm line
>180, <GEN2> line 172.
>
>I checked to variable and the %input{pA} and others were empty.
>
>Thanks into looking into this,
>
>	-Heikki
>
>
>On Tue, 2003-11-11 at 16:18, Stefan Kirov wrote:
>  
>
>>Yup:-)), actually you can supply two types of input data - arrays or 
>>strings:
>>-pA=>\@pA, where @pA=(0.7,0.3,0.5,0.1,0.9)
>>or
>>-pA=>$pA, where $pA='73519'
>>So the lines you commented out are necessary to parse the string input, 
>>however I guess it is my mistake somewhere if the input type is not 
>>recognized correctly and SiteMatrix is trying to parse arrays as 
>>strings. Could you send me the warnings you are getting? I am runnning 
>>5.8.0 and did not get any.
>>Thanks!
>>Stefan
>>
>>Heikki Lehvaslaiho wrote:
>>
>>    
>>
>>>Stefan,
>>>
>>>t/psm.t was printing out warnings originating from
>>>Bio::Matrix::PSM::SiteMatrix (under perl 5.8.2). I commented out some
>>>lines in the constructor that were trying to split non-existing hash
>>>values. All tests pass. Could you check that I did not do anything
>>>stupid.
>>>
>>>Cheers,
>>>	-Heikki
>>>
>>>
>>> 
>>>
>>>      
>>>

-- 
Stefan Kirov, Ph.D.
University of Tennessee/Oak Ridge National Laboratory
1060 Commerce Park, Oak Ridge
TN 37830-8026
USA
tel +865 576 5120
fax +865 241 1965
e-mail: skirov at utk.edu
sao at ornl.gov