whole genome seq for Gazelle - assemble mitochondria

Question

[m90]

Hello there,

I ran the Mitoz tool to extract all raw data about of mitochondria from DNA read about 90 GB and assemble it, but I got an error message that I couldn't understand. Here's the code that I used:

mitoz all \
  --outprefix Idmi \
  --thread_number 16 \
  --clade Chordata \
  --genetic_code 2 \
  --species_name "Gazella arabica" \
  --fq1 V350132114_L04_4_1.fq.gz \
  --fq2 V350132114_L04_4_2.fq.gz \
  --fastq_read_length 100 \
  --data_size_for_mt_assembly 0,0 \
  --assembler megahit \
  --kmers_megahit 59 79 99 \
  --memory 150 \
  --requiring_taxa Chordata

and here are the error messages:

024-07-02 13:22:04,493 - mitoz.utility.utility - ERROR - 
Error occured when running command:
megahit  --out-dir ./megahit_out  --num-cpu-threads 16 --k-list 59,79,99 --memory 1500000000000.0 -1 /home/ubuntu/Idmi_reference/clean_data/Idmi.clean_R1.fq.gz -2 /home/ubuntu/Idmi_reference/clean_data/Idmi.clean_R2.fq.gz

2024-07-02 13:22:04,494 - mitoz.utility.utility - WARNING - 
Can not find assembly_file: /home/ubuntu/Idmi_reference/mt_assembly/megahit/megahit_out/final.contigs.fa

[GenoMax]

Can not find assembly_file: /home/ubuntu/Idmi_reference/mt_assembly/megahit/megahit_out/final.contigs.fa

Message seems to be clear. It is not able to find this file.

[m90]

What assembly file ? It always create his own files that it needs in running

[GenoMax]

Then the error must be upstream. Look through the complete mitoz log file to see what may be going wrong. Part you posted here does not seem to have usable information.

Answer 1

[colindaven]

It looks to me that megahit could not create an output file due to an internal crash, but the tool mitoz did not pass on an error.

Maybe the RAM you assigned is too small - is this 150 GB RAM ? --memory 150.

You can monitor RAM constantly for this process on the server using a tool like htop and see if the RAM usage is increasing towards 150GB and it suddenly crashes/exits. If this is the case, try increasing RAM for the tool if possible.

[m90]

I will try but for 90 GB data , how much ram i should put ? I running on server about 750 Gb ram