PBJelly error at mapping level

Question

[SJ Basu]

Hello,

I have Illumina PE and Pacbio reads from a bacteria...now I have assembled Illumina PE reads using velvet and want to perform combined assembly with velvet-assembled-scaffold and PacBio reads using PBJelly... Though I have gone through the installation process without much hassle, I am stuck at the mapping level without any idea for more than two days now...I get this error msg after executing "jelly.py mapping myProtocol.xml" : IOError: [Errno 2] No such file or directory: '/path/to/my/directory/data/pbjelly_out2/mapping/DRR030785-87.fastq.m4' I have put all the path on bashrc and blasr is also running fine...i followed the instructions from a similiar post in pbjelly forum...still coudnt make it work !!!

Does anyone has experience with this tool ????? any help will be highly appreciated

[Jean-Karim Heriche]

Check similar issues on the PBSuite forum: here and here.

[SJ Basu]

First...@Jean-Karim...Thankyou so much for the time and reply...but those links were what I meant when I said "i followed the instructions from a similiar post in pbjelly forum" Secondly...This problem actually happens due to non-availability of m4 file (synonymous to SAM file but for blasr) so you have to run blasr from outside with -m 4 option and keep the output file in the path given in the error (along with the same file name given in error msg). That solves the issue.

[Jean-Karim Heriche]

Then the question is why is blasr not running from the main pipeline ? If it's in the path, maybe the pipeline is run as another user. Also I saw somewhere that the pipeline scripts run with /bin/sh so may be this is not pointing to the shell you're using.

Answer 1

[BenNab]

I had the same problem and I fixed it by editing m4pie.py and Stages.py.

You should change the blasr argument from the "-" prefix to "--" throughout. For example, -sa and -out needs to be changed to --sa and --out. Also, template files need to be changed accordingly and -minPctIdentity should be remaned to --minPctSimilarity

Answer 2

[caizexi123]

Hi,

The problem is caused by the parameter differences between old blasr and new blasr. As BenNab said, you need to modify the m4pie.py and Stages.py by changing all the parameters the authors give to blasr, except "-m", you need to change "-" to "--", but leave "-500", this is for the negative score for blasr. And -minPctIdentity should be remaned to --minPctSimilarity. But modified these two script is not enough. Because assembly stage, PBjelly use blasr again, so you also need to modify Assembly.py scripts as above. Besides, PBjelly do not accept any "_" in the names of scaffolds or reads.

[ruby]

Would you possibly be willing to share your copy of each of these files? I have attempted this correction and am still having the issue, and would like to verify that I have done this correctly.

[alexis.groppi]

Here are the corrected python scripts. PBsuite_modified_scripts.zip Don't forget to modify also the jellyprotocol.xml file The first step is working (mapping) but for an unexplained reason, blasr doesn't take the nproc argument. It works only with 1 proc... making this step very long So be patient ;)

[11228057]

Hi

Could you please share your corrected PBsuite scripts again? We met the same problems.

Thank you very much.

[zrlewis]

@alexis.groppi I'd also be interested in trying your modified scripts. I tried the modifications myself and must have missed a flag somewhere.

[scnkyjzsxy]

@alexis.groppi I'd also be interested in trying your modified scripts. But now the download page can not be reached. Can you provide it again, thanks.

[alexis.groppi]

Sorry for the delay here are the scripts : "PBsuite_modified_scripts.zip"

[R.Menafra]

@alexis, would you be so kind to add the scripts again? The link is not available anymore and unfortunately I'm struggling as well with the same problems! Thanks

[lagartija]

Hi, sorry to bother, could you please share with us again your modifies scripts ? I also tried Jelly2 and it comes with similar problems... would be so useful if it worked

[cgjosephlee]

Try installing from the latest commit svn checkout https://svn.code.sf.net/p/pb-jelly/code-0/trunk pb-jelly-code-0, tested on blasr 5.3 (8d086d7). Or checkout my repo https://github.com/cgjosephlee/PBJelly.

Answer 3

[11228057]

Hi

Could you please share your corrected PBsuite scripts again? We met the same problems.

Thank you very much.

Answer 4

[SJ Basu]

You are right...I too did see it somewhere in my error file but I can't trace back that error file...now my error files shows mapping_chunk0.out: ERROR: sa is not a valid option.
mapping_chunk0.err:IOError: [Errno 2] No such file or directory: '/path/to/my/directory/data/pbjelly_out2/mapping/sample85-87.fastq.m4' I am like stuck on this forever....i tried putting variables to bashrc, bash_profile, environment file still its not working...even the dependencies are up to date !!!! How do we solve this ?????? :(

[Surya Saha]

This is due to the -- prefix for parameters in blasr 5.1 instead of -. Please see https://sourceforge.net/p/pb-jelly/discussion/pbjtiks/thread/7657e087/

You can run blasr on the command line after modifying the parameter calls appropriately to generate the .m4 file. And yes, this should be fixed in PBSuite.

[Seigfried]

I could not get the newest version of Blasr 5.2.3194382 to work with PBSuite_15.8.24. Same error.

Do you mean that we should run blasr separately on the command line? and then link the output to the next step of PBJelly? If so, we can use the Sawriter output generated from the the first Setup step right?

Something like blasr reads.fasta reference.fasta -sa reference.fasta.sa