Assembling fails at 'builder build' step

[sipesk]

Hello,

I'm trying to use metaWRAP to assemble a metagenomic data set (~80GB), and it cannot get past the assembly step. Here is the code and the log file:

$ metawrap assembly -t 14 -m 38657077982 -1 IC8_2_3728_read1.fastq -2 IC8_2_3728_read2.fastq -o ./IC8_output_unzipped

%% This is the ERROR message at the bottom of the log file

[ERROR] [cx1_seq2sdbg.cpp         : 540]: 24000000000 bytes is not enough for CX1 sorting, please set -m parameter to at least 38657077981
Error occurs when running "builder build" for k = 21; please refer to ./IC8_output_unzipped/megahit/log for detail
[Exit code 1]

Thanks for your time, Kaite

[ursky]

Like the error says, you are running out of RAM. Make sure to set it with the -m parameter in metawrap.

[sipesk]

However, in my command, I set the -m to be higher than what it says it needs, and it still gives me that error.

[ursky]

The -m papameter sets it in GB, so try setting -m 100. You setting such an impossibly large number might be messing with the program. Also, do you actually have enough RAM on your system?

[marcomeola]

On the same line I got following error:

Error occurs when running "sdbg_builder count/read2sdbg", please refer to ASSEMBLY_megahit/megahit/log for detail

in the log file:

[ERROR] [cx1_kmer_count.cpp       : 157]: 10000000000 bytes is not enough for CX1 sorting, please set -m parameter to at least 24421563490
Error occurs when running "sdbg_builder count/read2sdbg", please refer to ASSEMBLY_megahit/megahit/log for detail

while the execution command was:

metaWRAP assembly -m 50 -t 34 -1 ./*_1.fastq -2 ./*_2.fastq --megahit -o ASSEMBLY_megahit

Despite the 50GB free-mem given to use, read2sdbg uses only 10GB. Any alternative on how I could increase mem?

[marcomeola]

[SOLVED]: The problem lies in the ASSEMBLY_megahit/megahit/opt.txt file, which records the parameters of your first try. If you need to increase threads or mem you better delete the complete folder or that particular file at least. Otherwise it will pick up again the old parameters ignoring the newer ones