#17 perl lib version differences causing error

[Wendy Phillips]

I will start with a many thanks for your help!

I've been beating myself over the head with this problem for a long time and am finally coming to you all for help. I am a computing novice, so please talk to me like a computing toddler.

I have installed the latest wgs version (8.2) as of Decmember 2014. This problem occurs when running pacBioToCA or PBcR. It seems the perl versions that the wgs machinery wants to use is in conflict with the version the smrtanalysis wants to use. By the way, I do all this work on a university server system. I have wgs installed in my local directory and can make modifications to the programs, but the smrtanalysis is installed in a public directory and I can't do any modifications to it.

My initial questions are: how do I resolve this error?
Can I restart the pacBioToCA or PBcR just before where these errors halted the run?
If so, what command would I use to restart the run?

Here is the error message I get in a file called 1.err.

"Perl lib version (v5.8.8) doesn't match executable version (v5.10.1) at /local/cluster/smrtanalysis/install/smrtanalysis-2.2.0.133377/miscdeps/basesys/usr/lib64/perl5/5.8.8/x86_64-linux-thread-multi/Config.pm line 46.
Compilation failed in require at / /home/ / /local/bin/wgs-8.2/Linux-amd64/bin/convertToPBCNS line 49.
BEGIN failed--compilation aborted at / /home/ / /local/bin/wgs-8.2/Linux-amd64/bin/convertToPBCNS line 49."

The problem seems to occur chronologically right after runPartition.sh* is added in the sequence of files/directories written. For example, here are files written on either side of the error occurrence.

lrwxrwxrwx 1 58 Dec 24 12:37 1.trim.qual -> /home/ / /genome/wgs//tempPBcorrectsmall12/1.qual
lrwxrwxrwx 1 59 Dec 24 12:37 1.trim.fasta -> /home/ / /genome/wgs//tempPBcorrectsmall12/1.fasta
-rw-r--r-- 1 0 Dec 24 12:37 1.trim.fastq
-rw-r--r-- 1 0 Dec 24 12:37 1.qual
-rw-r--r-- 1 0 Dec 24 12:37 1.success
-rw-r--r-- 1 81 Dec 24 12:37 1.lay.err
-rw-r--r-- 1 0 Dec 24 12:37 asm.1.log
-rw-r--r-- 1 454 Dec 24 12:37 1.err
-rwxr-xr-x 1 2.9K Dec 24 12:37 runPartition.sh*
-rw-r--r-- 1 0 Dec 24 12:37 asm.layout.success
-rw-r--r-- 1 11K Dec 24 12:37 asm.layout.err
-rw-r--r-- 1 2.7K Dec 24 12:35 asm.paired.list
-rw-r--r-- 1 6.1M Dec 24 12:35 asm.9.paired.olaps
-rw-r--r-- 1 4.1M Dec 24 12:35 asm.1.paired.olaps
-rw-r--r-- 1 3.9M Dec 24 12:35 asm.2.paired.olaps
-rw-r--r-- 1 3.8M Dec 24 12:35 asm.4.paired.olaps
-rw-r--r-- 1 3.4M Dec 24 12:35 asm.3.paired.olaps

Here is another detail that might be relevant. One of our biocomputing consultants helped me overcome an inability of the wgs programs to find blasr by submiting my job to the queue in the form of the following script:

"#!/usr/bin/env bash

source /local/cluster/smrtanalysis/install/smrtanalysis-2.2.0.133377/etc/setup.sh

pacBioToCA -libraryname PBcorrect8 -s spec5.file pacbio.fastq -length 500 -threads 16 -genomeSize 100000000 illuminas.frg > PBcorrect8try.out 2>&1"

In a different version of wgs I would get a similar perl version error, and the consultant had me resolve it by modifying the top line of pacBioToCA to this, which my current pacBioToCA had but I can't remember if it came like this or I modified it:
"#!/usr/bin/env perl"

I notice that the top line of convertToPBCNS is:
"#!/usr/bin/perl"
which used to be what pacBioToCA had. Do I need to go through all the wgs programs and make sure the first line is
"#!/usr/bin/env perl"
or is there a much simpler fix than all this?

Sorry for the long worded question.
Again, many thanks!
Wendy

[Sergey Koren]

Hi,

This looks like there is a conflict between the perl version in /usr/bin/perl which is newer than the the libraries being loaded (which are coming from SMRTportal). Is this happening after you run:
source /local/cluster/smrtanalysis/install/smrtanalysis-2.2.0.133377/etc/setup.sh

If so, then yes the solution would be to update the first line of two scripts:
convertToPBCNS
addCNSToStore
from #!/usr/bin/perl to #!/usr/bin/env perl

As far as resuming, you can run the following commands:
cd /home/genome/wgs/tempPBcorrectsmall12
rm runCorrection.sh
rm runPartition.sh
rm *.success
cd -

and re-launch the pipeline with pacBioToCA as you did previously.

[Wendy Phillips]

Sergey,

Thanks so much for the response! I actually already did modify the first line in convertToPBCNS yesterday and started a trial run of pacBioToCA, which seemed to complete successfully. I will do the same with addCNSToStore. I am now trying to optimize pacBioToCA parameters to get a higher resulting coverage of the corrected PacBio reads before continuing on to a full PBcR run. I will try the commands to resume the run on a big data set I had started previously.

Thanks again!
Wendy

[Sergey Koren]

CA 8.3 uses /usr/bin/env perl instead of /usr/bin/perl in all PBcR perl scripts.