ERROR:PBcR Aborted (terminate called after throwing an instance of 'std::bad_alloc')

Question

[349000955]

Hi all: i assemble a genome size 1.8 Gb with 40X pacbio reads and 30X NGS reads.when the program run at overlapStoreBuild it give a ERROR: Failed with signal ABRT,the program Aborted core dumped and exit.I look into the asm.ovlStore.BUILDING file there is a file about size 2.4 T while other file are about 180Gb.

/software/wgs-8.3rc2/Linux-amd64/bin/PBcR -length 1000 -partitions 200 -libraryname AA2 -threads 50 -genomeSize 1800000000 -s pb.spec -fastq ../00_data/AA2_Pcbio.fastq ../04_data4NGS/flashout/InsertSize_250.notCombined.frg ../04_data4NGS/flashout/InsertSize_346.notCombined.frg ../04_data4NGS/flashout/InsertSize_350.notCombined.frg ../04_data4NGS/flashout/InsertSize_500.notCombined.frg

i use flash to flash the NGS pe reads and use fastq2ca to generate NGS frg.Can you give some suggestion about my question.

[349000955]

Thank you suggestion rhall,but my boss require assemble the genome with NGS data and pacbio data together,and he think it will give a better result with 30X NGS data added.

[GenoMax]

@rhall works at PacBio and it would be wise to heed his advice.
Convince your boss that you want to try @rhall's way first and if that does not work do the thing your boss wants.

[349000955]

OK,thank you and i will have a try.Also i assembly those pacbio data to generate a genome size 1.45 Gb contig N50 197 kb,will canu have a better result in genome size and N50 compare with falcon?

[GenoMax]

Can't say. You won't know until you try canu.

Answer 1

[rhall]

I would suggest assembling the 40x pacbio alone (http://canu.readthedocs.io/en/stable/) then using the other data as a validation. The hybrid approach is going to be very compute intensive, and will likely not give any better results.