[Bioperl-l] codon usage
Jim Hu
jimhu at tamu.edu
Mon May 7 23:08:36 UTC 2012
Actually, I was looking at the wrong method for the topic of codon usage. Count codons doesn't use the regex; it just peels off triplets and uses a hash to track the counts.
I was looking at count_monomers, which uses this:
...
# For each letter, count the number of times it appears in
# the sequence
LETTER:
foreach $element (@$alphabet) {
# skip terminator symbol which may confuse regex
next LETTER if $element eq '*';
$count{$element} = ( $seqstring =~ s/$element/$element/g);
}
if ($_is_instance) {
$self->{'_monomer_count'} =\% count; # Save in case called again later
}
return\% count;
In this case the theoretical problem is that the regex engine is calling the $seqstring repeatedly, so that even if the engine is using a state machine, its repeating the scan of the sequence for each symbol in the alphabet. This is still likely to be way faster than an interpreted state machine, since the elements are just single characters and there are not many of them.
As I said in the first email, a state machine written for the interpreter running a more efficient algorthm probably would still lose to beat a built-in engine running a less efficient algorithm. And being able to stream instead of holding the whole sequence in memory is not a concern now, as it was when we had 4K of RAM.
JH
On May 7, 2012, at 2:15 PM, Joel Martin wrote:
> I think you're under-appreciating the perl regex engine, it operates
> as a state machine not by matching all occurrences every time. Here
> is a nice excerpt from pro perl parsing
> http://www.devshed.com/c/a/Perl/Parsing-and-Regular-Expression-Basics/2/
>
> Joel
>
> On Mon, May 7, 2012 at 8:28 AM, Jim Hu <jimhu at tamu.edu> wrote:
>> I was looking at Bio::Tools::SeqStats. It reminds me of the very first bioinformatics program I worked on back when I was in grad school, sequencing was done by Maxam-Gilbert chemistry on gels with radioactive DNA (we were doing short reads, but we didn't call it that and the reads per run was something like 8). We were writing a program in Apple II basic to find restriction sites. Everyone in the group was doing this by putting the target sequence in a string variable and looking for the site as a substring match by whatever it was BASIC used. Loop over all the sites you are looking for. This strikes me as the equivalent of how Bio::Tools::SeqStats works, only with regexes.
>>
>> My roommate at the time, who was a math PhD student doing an MS in CompSci, pointed out to me that this would be more efficient using a discrete finite automaton algorithm, where each site we were looking for would be a state automaton. This has the advantage of being able to process the sequence as a stream. Back when we were working with computers with RAM measured in Kbytes, this was a big help. I'm not sure if it would be worth it today. The slow interpreted implementation of the state machines would likely lose to the fast internal implementation of the regex routines for sequence lengths we are looking at these days.
>>
>> But it might be interesting to compare.
>>
>> Jim
>>
>> On May 6, 2012, at 3:52 PM, Smithies, Russell wrote:
>>
>>> Or use Bio::Tools::SeqStats
>>> (this is straight from the perldocs http://search.cpan.org/~cjfields/BioPerl-1.6.901/Bio/Tools/SeqStats.pm )
>>>
>>> $seqobj = Bio::PrimarySeq->new(-seq => 'ACTGTGGCGTCAACTG',-alphabet => 'dna',-id => 'test');
>>> $seq_stats = Bio::Tools::SeqStats->new(-seq => $seqobj);
>>>
>>> $hash_ref = $seq_stats-> count_codons(); # for nucleic acid sequence
>>> foreach $base (sort keys %$hash_ref) {
>>> print "Number of codons of type ", $base, "= ", %$hash_ref->{$base},"\n";
>>> }
>>>
>>>
>>> --Russell
>>>
>>>
>>> -----Original Message-----
>>> From: bioperl-l-bounces at lists.open-bio.org [mailto:bioperl-l-bounces at lists.open-bio.org] On Behalf Of subarna thakur
>>> Sent: Saturday, 21 April 2012 3:00 p.m.
>>> To: bioperl-l at bioperl.org
>>> Subject: [Bioperl-l] codon usage
>>>
>>> I am writing a script for determining number of genes containing a particular codon. The codons are mentioned in a separate file. The output is coming all right for the first codon mentioned in the file but for the other codons , the script is not working. Please suggest the error in the script. The script is as follows ----
>>>
>>> #!/usr/bin/perl -w
>>>
>>> use Bio::SeqIO;
>>>
>>> $file2="table.txt";
>>>
>>> $codon=0;
>>>
>>> open OUT, ">out-test.txt" or die $!;
>>>
>>> $seqio_obj = Bio::SeqIO->new( -file => "gopi2.txt" , '-format' => 'Fasta');
>>>
>>> open( my $fh2, $file2 ) or die "$!";
>>>
>>> while( my $line = <$fh2> ){
>>>
>>> $acc=$line;
>>>
>>> chomp $acc;
>>>
>>> while ($seq1= $seqio_obj->next_seq){
>>>
>>> my @output = $seq1->id;
>>>
>>> my $string = $seq1->seq;
>>>
>>> $v=0;
>>>
>>> $l= length($string);
>>>
>>> $t=$l/3;
>>>
>>> $k=0;
>>>
>>> for ($i=1; $i <= $t; $i++){
>>>
>>> @array2 = substr($string, $k, 3);
>>>
>>> $k=$k+3;
>>>
>>> foreach $value (@array2)
>>>
>>> {
>>>
>>> if ($value eq "$acc")
>>>
>>> {
>>>
>>> print OUT " The sequence id is @output\n";
>>>
>>> print OUT "$acc codon found in position $i\n\n";
>>>
>>> $v=$v+1;
>>>
>>> }
>>>
>>> }
>>>
>>> }
>>>
>>> if ($v==0)
>>>
>>> {
>>>
>>> $h=0;
>>>
>>> }
>>>
>>> else
>>>
>>> {
>>>
>>> $h=1;
>>>
>>> }
>>>
>>> $codon=$codon+$h;
>>>
>>> }
>>>
>>> print OUT "Total number of sequences with $acc codon";
>>>
>>> print OUT "\t";
>>>
>>> print OUT $codon;
>>>
>>> }
>>>
>>> exit;
>>> _______________________________________________
>>> Bioperl-l mailing list
>>> Bioperl-l at lists.open-bio.org
>>> http://lists.open-bio.org/mailman/listinfo/bioperl-l
>>> =======================================================================
>>> Attention: The information contained in this message and/or attachments
>>> from AgResearch Limited is intended only for the persons or entities
>>> to which it is addressed and may contain confidential and/or privileged
>>> material. Any review, retransmission, dissemination or other use of, or
>>> taking of any action in reliance upon, this information by persons or
>>> entities other than the intended recipients is prohibited by AgResearch
>>> Limited. If you have received this message in error, please notify the
>>> sender immediately.
>>> =======================================================================
>>>
>>> _______________________________________________
>>> Bioperl-l mailing list
>>> Bioperl-l at lists.open-bio.org
>>> http://lists.open-bio.org/mailman/listinfo/bioperl-l
>>
>> =====================================
>> Jim Hu
>> Professor
>> Dept. of Biochemistry and Biophysics
>> 2128 TAMU
>> Texas A&M Univ.
>> College Station, TX 77843-2128
>> 979-862-4054
>>
>>
>>
>> _______________________________________________
>> Bioperl-l mailing list
>> Bioperl-l at lists.open-bio.org
>> http://lists.open-bio.org/mailman/listinfo/bioperl-l
=====================================
Jim Hu
Professor
Dept. of Biochemistry and Biophysics
2128 TAMU
Texas A&M Univ.
College Station, TX 77843-2128
979-862-4054
More information about the Bioperl-l
mailing list