Skip reads correction step in metaSPAdes

[mchialva]

Hi,
I'm trying to assemble complex soil metagenomes (2x150, Novaseq) at a relatively high sequencing depth using metaSPAdes and I was wondering if it can be worth to skip the reads corrections step. I read somewhere in previous Q&A here that reads correction in large datasets, beside increasing computational time, tend to collapse similar sequences. I tried both option on a test metagenome and found that without reads correction the contig number increased of four times. What would you recommend on this point?
Thanks

[asl]

Well, absolute contig number frankly speaking does not mean anything. As SPAdes does not do any filtering of the output, the extra contigs might be just some uncorrected stuff.

I would recommend you to analyze the results in terms of the contents, length, etc.