Hey, I have Illumina 150 PE sequence reads of water/soil samples. It is a shotgun metagenomics and there are 20 samples and each have about 10 millions reads. I am trying figure the reads assembly method. I have tried velvet, idba_db and SPAdes. Velvet seems taking a long time (>48hours for one sample). SPAdes also take a long time and the N50 is not so great, which is 240bps using K=55). IDBA just keep crashing and based on the discussion online I think it has to do with my reads being longer than what the software was designed.

So has anyone had any success in de novo assembling of 150 PE Illumina reads of environmental samples? If so, please share. Thanks.