search

DaehwanKimLab / hisat2

Graph-based alignment (Hierarchical Graph FM index)
GNU General Public License v3.0
464 stars 113 forks source link

hisat-3n-table #306

Closed pillailab closed 2 years ago

pillailab commented 2 years ago

I am trying to run hisat-3n-table on the sam file after hisat-3n alignment ( the sam file was about 50 Gb). I provided 20 cores with 6 Gb each; the slurm job runs for over a day with no output and no error message. Here is the script: hisat-3n-table -p 20 \ --ref $genome \ --alignments "$b".hisat-3n.sam \ --output-name "$b".hisat-3n.table.tsv \ --base-change T,C Thank you for your help

Manoj

imzhangyun commented 2 years ago

Hello Manoj,

  1. What is your --ref format? Is it FASTA?
  2. Can you check the memory usage for hisat-3n-table?
  3. For high coverage reads SAM, hisat-3n-table may use over 10GB memory. Have you every try to use hisat-3n-table use only one file and large memory (~16GB)?
  4. Did you use --remove-chrname or --add-chrname option during hisat-3n alignment?

Best, Leo

pillailab commented 2 years ago

Hi Leo,

  1. yes, fasta format
  2. memory usage is very low (less than 1%)

Job ID: 4571720 Cluster: farnam User/Group: mp758/pillai State: TIMEOUT (exit code 0) Nodes: 1 Cores per node: 20 CPU Utilized: 1-11:32:02 CPU Efficiency: 7.40% of 20-00:07:20 core-walltime Job Wall-clock time: 1-00:00:22 Memory Utilized: 340.11 MB Memory Efficiency: 0.28% of 117.19 GB

  1. did you mean one core and large memory ? I will try that.
  2. Yes, i did use --add-chrname. below are the options i used: hisat-3n -x $index \ -p 20 \ -1 "$b".t.R1.fastq.gz \ -2 "$b".t.R2.fastq.gz \ --base-change T,C \ --add-chrname \ --rna-strandness FR \ --mp 4,2 \ -S "$b".hisat-3n.sam

Is that a problem to add chrrname ? Many thanks ! Manoj

imzhangyun commented 2 years ago

Hello Manoj,

Unfortunately current hisat-3n-table does not support the SAM file with you using --remove-chrname or --add-chrname during alignment. Because your reference is "18" and your SAM file shows "chr18", hisat-3n-tablecannot find the correct chromosome.

I suggest you remove the --add-chrname in your current pipeline during alignment. If the "chr" prefix is very important for you, you may add "chr" prefix on your reference (fasta) file.

I will edit the hisat-3n-table and make it supports the SAM with --remove-chrname or --add-chrname soon.

Best, Leo

imzhangyun commented 2 years ago

Hey Manoj,

Good news! I just add the option --added-chrname and --removed-chrname for hisat-3n-table.

Please use git pull and make to update the newest hisat-3n-table. Please let me know if you have any other question.

Best, Leo

pillailab commented 2 years ago

Thanks Leo, I will try that. I did realign without the --add-chrname and it seems to have done the trick. Best, Manoj