I'm planning my first metagenomic experiment and hope to get some advice/comments before I start working on it.
We are trying to explore the antimicrobial properties of a plant and the idea here is to investigate the "killing effect" of the plant on the oral metagenomes. We are going to do shotgun metagenomes at 3 different time points after eating the plant. Initially we were thinking of depleting the human DNA before sequencing. However, the sequencing company has advised us to instead increase the sequencing depth by at least 10x to allow for better titration of the microbial species. Of course, this will push up our cost quite significantly. Is this really necessary?
In addition to the shotgun metagenomic studies, we are also thinking of looking at perhaps another 50 cases by doing 16s rRNA metagenomic sequencing instead. I know the resolution is going to be less than the shotgun approach but will the number of cases give us the statistical confidence needed?
Another concern we have is how much difference can we detect in each of the time point. I guess not all the DNA from the "killed" microbes is going to be totally degraded right?
Any comments will be highly appreciated. Thanks.
In addition, what sort of read length should we consider for both the whole genome shortgun and 16s metagenomics? Thanks.
I would highly recommend looking into http://onecodex.com/ as a down stream analysis tool for profiling your sequencing results.