Issues assembling using ABySS on HTC

Question

[kingcohn]

Hello ABySSers, I am attempting to assemble a 30 mega base genome using AbySS on my university’s HTC. The test-data was able to assemble without error, but when I ran my samples I received an mpirun error:

bin/abyss-pe np=8 k=61 name=abyssk61 in='CP3.1121.fastq CP3.1222.fastq' /mnt/gluster/chtc/openmpi-1.6.4/bin/mpirun -np 8 ABYSS-P -k61 -q3 --coverage-hist=coverage.hist -s abyssk61-bubbles.fa -o abyssk61-1.fa CP3.1121.fastq CP3.1222.fastq mpirun was unable to launch the specified application as it could not find an executable: Executable: ABYSS-P Node: e288.chtc.wisc.edu while attempting to start process rank 0. 8 total processes failed to start make: [abyssk61-1.fa] Error 133

After reading on the Google Group, I loaded the proper modules using

module load compile/intel module load openmpi-1.6.4

and reconfigured using:

./configure –with-mpi=/mnt/gluster/chtc/openmpi-1.6.4 checking mpi.h usability... yes checking mpi.h presence... yes checking for mpi.h... yes checking for MPI_Init in -lmpi... yes However, the make command failed with: make CDPATH="${ZSH_VERSION+.}:" && cd . && /bin/sh /var/lib/condor/execute/slot1/dir_747618/abyss-1.9.0/missing aclocal-1.13 /var/lib/condor/execute/slot1/dir_747618/abyss-1.9.0/missing: line 81: aclocal-1.13: command not found

WARNING: 'aclocal-1.13' is missing on your system. You should only need it if you modified 'acinclude.m4' or 'configure.ac' or m4 files included by 'configure.ac'. The 'aclocal' program is part of the GNU Automake package: http://www.gnu.org/software/automake It also requires GNU Autoconf, GNU m4 and Perl in order to run: http://www.gnu.org/software/autoconf http://www.gnu.org/software/m4/ http://www.perl.org/ make: [aclocal.m4] Error 127

Even after I installed those dependencies with wget. Here is my current ABySS directory:

ABYSS ChangeLog config.sub DataLayer Graph LICENSE missing README.md aclocal.m4 CITATION.bib configure depcomp index.html LogKmerCount Overlap Scaffold AdjList CITATION.md configure.ac dialign index.html.1 m4 PairedDBG Sealer Align Common Consensus DistanceEst install-sh Makefile Parallel SimpleGraph Assembly config.guess COPYRIGHT doc KAligner Makefile.am ParseAligns stamp-h1 autoconf config.h CP3.1121.fastq doxygen.conf kmerprint Makefile.in PathOverlap test-driver automake config.h.in CP3.1222.fastq FilterGraph Konnector Map PopBubbles Unittest bin config.log DAssembler FMIndex Layout MergePaths README.css Bloom config.status DataBase GapFiller lib Misc README.html

Any suggestions would be appreciated, thank you!

[benv]

Hi Zachary,

Are you compiling from the git repository or from the source tarball? I would recommend using the source tarball, because in that case you won't require autoconf/automake. (If you are compiling from the git repository, you need autoconf and automake installed and you also need to run ./autogen.sh before ./configure).

It looks like ./configure is finding your MPI library correctly, so it should build ABYSS-Pfor you, once you get rid of the make errors.

[kingcohn]

Hey Benv,

So, I've been able to compile and install the software on my university's HTC system. However, after running for several days there are no outputs and the system cancels my jobs. I've run other assembly softwares on the network, spades, minia...etc. without this issue. Here is the script for assembly:

#!/bin/bash

#untar abyss and CD into bin directory with ./abyss.pe 
tar -xvzf binforabyss.tar.gz
cd bin #untarred Abyss binaries

. /etc/profile.d/modules.sh
module load openmpi-1.6.4 #for multithreading (used for np, but not used in this abyss-pe command)
ls

cp /mnt/gluster/zcohen3/CPthreeclean_1.fq.gz ./
cp /mnt/gluster/zcohen3/CPthreeclean_2.fq.gz ./
gunzip CPthreeclean_1.fq.gz #open forward PE data file
gunzip CPthreeclean_2.fq.gz #open reverse PE datat file
ls
mkdir abysstry 
pwd
./abyss-pe k=71 name=abysstry in='CPthreeclean_1.fq CPthreeclean_2.fq'

#cp output from working directory into gluster
cp -r abysstry /mnt/gluster/zcohen3
ls /mnt/gluster/zcohen3
rm -r abysstry

Answer 1

[kingcohn]

Hello Benv,

Thanks for your reply. Yes, I installed from the source.tar and used the most recent update (2.0.0) from yesterday 9/1. This update no longer requires the sqlite library and appears to have installed properly. Tangentially, do you have any suggestions for assembling low coverage (20x) illumina paired-end reads from two data sets/ versions? I concatenated the two forward and two reverse reads and am assembling using those fastq files, but is there a more efficient approach? Thanks again.

[benv]

Hi @kingcohn,

Glad to hear. No need to concatenate FASTQs prior to running abyss-pe. You can specify more than two FASTQs files per paired-end library (preferably ordered like: read1a.fq.gz read2a.fq.gz read1b.fq.gz read2b.fq.gz) or multiple paired-end libraries, if the two pairs of FASTQs have different fragment size distributions (i.e. they are from different DNA libraries). See https://github.com/bcgsc/abyss#assembling-multiple-libraries, for example.