Error While Running QuorUM

[mborche2]

I am trying to run HG-CoLoR for error correction of PacBio reads with 10x Linked-Reads (short reads). I installed QuorUM with bioconda. I got this error when running HG-CoLoR:

----- QuorUM -----
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LANG = "C.UTF-8"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
terminate called after throwing an instance of 'std::runtime_error'
what(): Hash is full
terminate called recursively
Creating the mer database failed. Most likely the size passed to the -s switch is too small. at /home/mborche2/miniconda3/envs/hg_color_env/bin/quorum line 155.

Any recommendations? Thanks a ton for putting out this software and helping others!

[morispi]

Hi,

Are you using short reads in FASTQ format, or in FASTA format? If you're using FASTA, that could be the issue: QuorUM requires the short reads to be in FASTQ format. PacBio reads, however, need to be FASTA.

Otherwhise, have you tried to launch QuorUM alone to see if you get the same error? Judging from the error messages, this sounds more like an error with your installation of QuorUM itself, or probably even of perl.

Another remark: I don't think QuorUM will produce any satisfying results with a 10x coverage of short reads. When developpng HG-CoLoR, I was having in mind coverages of around 50x of short Illumina reads. What are the specs of your Linked-Reads in terms of error rate / length? Maybe you could bypass the QuorUM step all together. I could provide you a script doing so.

Cheers,
Pierre

[mborche2]

Thanks for the extremely fast response! I am using a FASTA for the long reads and a FASTQ for the short read. I just tried QuorUM on its own and it isn't working on my short reads.

Also, the short reads are from the company 10x Genomics and the coverage is around 70-80x. I think the reads are still produced by an Illumina sequencer though. It looks like QuorUM only error corrects the short reads right? I'm more interested in error correcting the PacBio reads. If bypassing QuorUM is an option that would be awesome!

[morispi]

Can you show me what QuorUM outputs when you run it on its own on your short reads? I never tried installing it through bioconda, so maybe something wrong happened during the installation.

Ooh, my bad, I thought you were using a 10x coverage of short reads. Didn't get you were talking about the 10x company. Everything should be fine with QuorUM then, indeed, as it only requires a relatively deep coverage in order to be able to perform correction.

Bypassing QuorUM is an option, but that would decrease the quality of the PacBio reads correction, as a graph is built from the short reads, and that this graph needs to be as accurate as possible. Maybe you could use another short read corrector though? Only used QuorUM for convenience, as it is fast and accurate.

P

[Yunxia-li]

Can you show me what QuorUM outputs when you run it on its own on your short reads? I never tried installing it through bioconda, so maybe something wrong happened during the installation.

Ooh, my bad, I thought you were using a 10x coverage of short reads. Didn't get you were talking about the 10x company. Everything should be fine with QuorUM then, indeed, as it only requires a relatively deep coverage in order to be able to perform correction.

Bypassing QuorUM is an option, but that would decrease the quality of the PacBio reads correction, as a graph is built from the short reads, and that this graph needs to be as accurate as possible. Maybe you could use another short read corrector though? Only used QuorUM for convenience, as it is fast and accurate.

P
Hi:
I have install HG-CoLoR through conda, also QuorUM.
But when I ran: HG-CoLoR --longreads long.fasta --shortreads fake.fq -S 3 -n 30 --out output -K 101.
But in the _Removing short reads containing weak K-mers_process, it said :
/anaconda3/bin/bin/filterShortReads.py: No such file or directory
So , does it a install problem or others?

Hope for your reply.
Yun

[morispi]

@Yunxia-li

Hi,

Unfortunately, I do not maintain the conda version of HG-CoLoR myself, I only maintain the GitHub. Maybe you could try installing from GitHub? The install should be pretty easy, running ./install.sh should be enough.

Most probably, the script "filterShortReads" that HG-CoLoR is looking for has a different name, or even is not present, in your /anaconda3/bin/bin/ folder. Maybe you could just copy it from the GitHub repo into this folder?

Hope that helps.

Cheers,
Pierre