CREST structural variation software error

[bubfranks]

Hi all, I'm getting an error using CREST's example pipeline for structural variation detection

$CREST.pl -f tumor.bam.cover -d tumor.bam -g germline.bam --ref_genome \
        hg18.fa -t /genome/hg18.2bit -r chr1
Use of uninitialized value $tid in array element at ./CREST.pl line 338.

It's not obvious to me from the code how $tid failed to initialize, and my search for similar errors came up empty. Before I try to isolate the problem I thought I would raise the error here to see if anyone else has seen it before.

Thanks!

[bubfranks]

Hm, perhaps the authors don't hang out here. I'll send a message directly. If I get a resolution I'll post here.

Meanwhile, any CREST success stories?

[flipwell]

I have run it and got a few of those errors. It still completed though and I've compared the results with what I found with BreakDancer - didn't find anything new, BreakDancer picked up around 80% of CREST results and the others had been picked up by Freec. Best thing was the list was much shorter than BreakDancer, of the 30ish predictions only 2 look a bit dubious in the mappings

[bubfranks]

Oh! the error I saw only occurred when I attempted to use the -r option. Without an -r parameter everything worked fine, cool.

But! It only works for paired-end input in my hands. You're supposed to be able to use CREST in --nopaired mode, but when I attempt --nopaired mode extractSClip.pl gives me a torrent of errors like:

Use of uninitialized value $sdna in substr at /usr/local/lib64/perl5/Bio/DB/Bam/AlignWrapper.pm line 243.

extractSClip does provide some output, in spite of the errors. But, when extractSClip's output is supplied to CREST.pl, CREST.pl returns a null file for xxx.predSV.txt and another mighty stream to STDERR.

So, it seems my issue remains. Has anyone successfully used CREST for unpaired SV analysis?

[bubfranks]

hm, so it turned out that I misinterpreted the instructions. CREST instructions clearly state you need to use the same genome file in the CREST pipeline as you used in the initial mapping, but I took that to mean it was OK as long as the assemblies were the same. After using the exact same hg18.fasta file in BWA and CREST then the program started to work for me.

[aquinom85]

I'm trying as well but getting a different error than you, regarding the BLAT server.

/Crest$ CREST.pl -f tumor.bam.cover -d tumor.bam -g germline.bam --ref_genome /home/aquinom/hg18.fa -t /home/aquinom/hg18.2bit --2bitdir /home/aquinom
Replacement list is longer than search list at /usr/local/share/perl/5.12.4/Bio/Range.pm line 251.
4 125893227 + 5
getaddrinfo() error on hostName=sjblat: No address associated with hostname

Connection timed out
Sorry, the BLAT/iPCR server seems to be down. Please try again later.
10 66301865 - 9
getaddrinfo() error on hostName=sjblat: No address associated with hostname

Connection timed out
Sorry, the BLAT/iPCR server seems to be down. Please try again later.
10 66301858 + 4
getaddrinfo() error on hostName=sjblat: No address associated with hostname

Connection timed out


I have gfServer running with the command: gfServer start localhost 3503 ~/hg18.2bit
and tested that it works using gfClient. Anyone know what I need to change to get this working?

[aquinom85]

Fixed issue, had to change the hardcoded values in the perl script.

[shruti]

Hi,
I am trying to use the new version of CREST program, CREST1-0-1. But I'm having some difficulties, I have tried to do some changes and managed to eliminate some of the errors.

Let me know if anyone is having the same issues..

1. on line 539 "my ($rs, $re) = ($m>0 ? $a->[$m-1][0] : $a->[0][0], $m<scalar(@{$a})-1 ? $a->[$m+1][0] : $a->[-1][0];" , the program was giving error for $m, as it was not declared
below is the subroutine

sub bin_search {
my ($a, $p) = @_;
my ($s, $e) = (0, scalar(@{$a})-1);
# my $m = int( ($s + $e)/2);
while(1) {
return $s if($a->[$s][0] >= $p);
return $e if($a->[$e][0] <= $p);
# return $m if($a->[$m][0] == $p || ($a->[$m-1][0] < $p && $a->[$m+1][0] > $p));
my ($rs, $re) = ($m>0 ? $a->[$m-1][0] : $a->[0][0], $m<scalar(@{$a})-1 ? $a->[$m+1][0] : $a->[-1][0];
return $m if($a->[$m][0] == $p || ($rs < $p && $re > $p));

if($a->[$m][0] > $p) {
$e = $m;
}
else {
$s = $m;
}
$m = int( ($s+$e)/2 );
}
}
I had then uncommented the line 534 "my $m = int( ($s + $e)/2);"

2. After this I got another error in the line 539 "my ($rs, $re) = ($m>0 ? $a->[$m-1][0] : $a->[0][0], $m<scalar(@{$a})-1 ? $a->[$m+1][0] : $a->[-1][0];" , syntax error at ];"
I then added a ) at the end and changed it to my ($rs, $re) = ($m>0 ? $a->[$m-1][0] : $a->[0][0], $m<scalar(@{$a})-1 ? $a->[$m+1][0] : $a->[-1][0];" , syntax error at ]);

3. The program now seems to work, but I'm getting another error which I'm not able to understand
Can't open /tmp/150008.1.all.q/DmcqZ93HGj/Q4KaHbp14u.fa.cap.ace:No such file or directory at /home/ssinha/sw/CREST1-0-1/CREST.pl line 595

Reading through the code I didn't understand where to change for the tmp folder

[tahamasoodi]

Hi Shruti,

I'm also facing the same error. Did you solve this error, please let me know.

can't open /tmp/aPNssgMFN0/QARvoJZlvh.fa.cap.contigs.clip.fa.psl.sorted:No such file or directory at ./CREST.pl line 606

[shruti]

Yes I managed to run it.. this was to do with some permission problem.
so I changed the path of the output to a location where I have full permission.

[dakin]

I'm also getting a similar error message when running CREST.pl:

Can't open /tmp/7iYXT0zlo_/e9TZ74LVc9.fa.cap.ace:No such file or directory at /crest/1.0/bin/CREST.pl line 593.

Update: It is running now that our gfServer is running and as long as I use the exact same path to start BLAT server on the .2bit file as I then use to return the prompt and start CREST.pl, similar to what other users report.