[maker-devel] substr outside of string

Michael Nuhn mnuhn at ebi.ac.uk
Wed May 1 05:38:52 MDT 2013


Hello!

I have run maker with est and rna seq data to create a training set for 
SNAP. Then I trained SNAP and added the hmm to the snaphmm option and 
reran maker.

Maker is giving me error messages like this:

"
setting up GFF3 output and fasta chunks
doing repeat masking
re reading repeat masker report.
<Name of file>
substr outside of string at <Path to 
maker>/maker/2.27/maker/bin/../lib/repeat_mask_seq.pm line 140
.
--> rank=NA, hostname=ebi-209.ebi.ac.uk
"

The line from which this error message originates is:

	substr($$seq, $b -1 , $l, "$replace"x$l);

After getting these error messages I replaced it with

       eval {
	substr($$seq, $b -1 , $l, "$replace"x$l);
       };
       if ($@) {
	use Carp;
	use Data::Dumper;
	confess(
	    $@
	    . "\n\n"
	    . Dumper($p)
	    . "\n\n"
	    . "Length of sequence: " . (length $$seq)
	);
       }

After that I got this:

$VAR1 = [
           98926,
           99033
         ];


Length of sequence: 98686 at <Path to 
maker>/maker/2.27/maker/bin/../lib/repeat_mask_seq.pm line 14
5

I have not changed the genome file.

I'm also concerned with the reported length of 98686, because I have a 
list of all sequences in the file and their lengths, and none of them 
has a length of 98686 bp. The sequences with the closest lengths are these:

98367 LSalAtl2s1200
98438 LSalAtl2s1473
98776 LSalAtl2s1613
98876 LSalAtl2s1199

so they are not even close.

$$seq is a sequence as a string, when I print it.

Sometimes maker prints a message like this:

"
--Next Contig--

Processing run.log file...
#---------------------------------------------------------------------
Now retrying the contig!!
SeqID: LSalAtl2s63
Length: 3997709
Tries: 5!!
#---------------------------------------------------------------------
"

But according to my list, which I generated from the exact same file 
that maker has in genome_file option, the length of that sequence is 
1169407.

Any idea, why I am getting these problems and what to do about them?

Cheers,
Michael.




More information about the maker-devel mailing list