Shallow shotgun metagenomics

[harbortc]

Hi Rahel31,
I am in a similar situation. I'm considering a shallow shotgun sequencing for several hundred fecal samples (from mice). Planning to try the Hackflex protocol and mapping to a mouse gut-specific metagenome database. Hoping to get started on some initial library prep and sequencing soon.

I would be happy to hear what you decided on, or if you have any results yet on Hackflex performance with a stool metagenome sample!

Good luck

[Rahel31]

Hi,
I am trying to switch from 16S to shotgun metagenomics, but as the number of samples is quite high (several hundreds of human stool samples), then deep sequencing is a bit problematic price wise. I have been reading up on shallow shotgun sequencing - basically like the deep one, but you pool more samples and aim for lower number of reads per sample (half a million reads in some studies, a bit more in others). A bioinformatician I talked with said that now it should be quite reasonable as the databases for human gut microbiome are becoming rather complete, so that you don't need to assemble the genomes de novo, but could map to the genomes in existing database. Then again, a company doing metagenomics as service said that the clients are often requesting shallow shotgun metagenomics, but then are unhappy with the results...
We are aiming at having a species, hopefully sub-species resolution and as well to have functional information on the gut microbiome. Not looking at exhaustively specific genes, like AMR-genes etc...

I would really appreciate if you could you please share your experience with shallow shotgun metagenomic sequencing! What limitations did you encounter etc.
Another option would be to cut down on library preparation costs, for example by using the Hackflex protocol, but as we are not experienced in shotgun library preparation, then maybe this is more dangerous path to go down with.
Thank you in advance!