Open EorgeKit opened 2 weeks ago
Hi!
You need to add -p
to your restart command.
SqueezeMeta.pl /home/maloo/SW_sample --restart -step 4
should be
SqueezeMeta.pl -p /home/maloo/SW_sample --restart -step 4
Also you can not use more than one node for the same sample. You should instead run several samples in parallel, each in one node. The sequential
mode can not achieve this, so you will need to
coassembly
mode for this (though it will not really do a coassembly since you will only have one sample per job).Hello Sequential projects must be restarted individually, specifying the project name, as Fernando said. Besides, I would recommend updating to 1.6.4, that will be released very shortly, where several issues were corrected. Best, J
Hello Sequential projects must be restarted individually, specifying the project name, as Fernando said. Besides, I would recommend updating to 1.6.4, that will be released very shortly, where several issues were corrected. Best, J
Oh yeah thanks, I managed to deploy the restart mode.
Hi!
You need to add
-p
to your restart command.
SqueezeMeta.pl /home/maloo/SW_sample --restart -step 4
should beSqueezeMeta.pl -p /home/maloo/SW_sample --restart -step 4
Also you can not use more than one node for the same sample. You should instead run several samples in parallel, each in one node. The
sequential
mode can not achieve this, so you will need to
- Make a different samples file for each sample, containing only the fwd and rev fastq files for that sample
- Launch adifferent SqueezeMeta job for each sample. You can use the
coassembly
mode for this (though it will not really do a coassembly since you will only have one sample per job).
Hi, thanks a lot , I manage to get a hold of a big memory node which kind took care of the problem
Hi there again, I managed to run all my samples using the sequential mode after getting the bigmemory
node.
Particularly I am ONLY interested in the cyanobacteria present in the three environmental samples. My proposed method is, after I am done with the assembly and all the steps, I mine all the contigs that classify as cyanobacteria and reanalyze their functional and taxonomy profiles independently. Do you think this is the right way or can you advise a better one?
Thank you in advance
That's ok. The subset functions in SQMtools are devised to help you on that Best,
Perfect , Thanks
Hi there once again,
One of my sample had a low mappingstat so decided to use the sqm_reads.pl
script.
Code:
sqm_reads.pl \
-p SW_sample_reads_based \
-s $SAMPLE_FILES/SW.samples \
-f $RAW_DATA_DIR/SW \
-t 52 -b 100
Project directory structure:
ls SW_sample_reads_based/
creator.txt SW_sample_reads_based.out.allreads SW_sample_reads_based.out.allreads.mcount
methods.txt SW_sample_reads_based.out.allreads.funcog SW_sample_reads_based.out.mappingstat
SW_sample SW_sample_reads_based.out.allreads.funkegg
When trying to run sqm2pavian.pl
script:
Code:
sqm2pavian.pl SW_sample_reads_based/
You can now use these data in your R pavian or in the web app https://fbreitwieser.shinyapps.io/pavian
But no pavian file is produced, it is worth noting that sqm2pavian.pl
produces a pavian file for the Squeezemeta.pl
runs .
Please advise on what might be the issue.
Best, G
Hi @fpusan @jtamames @ggnatalia I am using squeezemeta to analyse my samples individually using the sequential mode. Thing is I am using a cluster that uses PBS job sceduler. I request a lot of nodes for each sample and they are assigned to me. However despite all that, almost everytime a job is killed during a diamond process despite me playing with different
-b
values.Each node has about 120GB of RAM and I request 6 nodes. That makes 720GB of RAM to be available per sample run. Dividing by 8 that makes -b value of about 90. What I discovered is that only one node is utilized during the diamond process for some reasons and all the others are left free. So I tried the automatic choice of block size by not specifying the block size and it manage to calculate that the node it works on has about 109GB of RAM so it assigns about
-b 11
and this makes diamond to run until step6 which involves the LCA algorithm and then it dies again. At this point I don't know what the problem is.All attempts to restart the run with
--restart
option have proven futile as I keep getting the error despite the fact that the file is in the project directory with read permissions:File availability:
My original code is :
Here is the syslog output:
Progress file output:
My nohup output error file:
My nohup output normal file(basically local syslog):