separating error correction for each fastq then combining at the end

[johnathanlo]

Hi there,

We are running into some memory issues while running hifiasm, probably due to the large amount of data. Our genome is around 2.2 Gb sequenced to about 150x coverage on 3 cells. This means there are 3 fastq files around 220 GB each. Unfortunately we are mostly limited to compute nodes with < 500 GB memory and the memory usage for the error correction step seems to be well in excess of 500 GB.

FYI, read quality looks good:

Would it be possible to separately generate error corrected reads for each batch of raw reads and then recombine them for assembly? Would you advise against this?

Thanks,
John