Hi Amanda,

are you talking about metagenomics in the sense of reconstructing genomes or are you talking about 16S rRNA gene amplicon sequencing? Just to clarify the question...

Alex

Just 16S rRNA amplicon sequencing - no genome reconstruction plans!

Hi Amanda--our lab has a lot of experience profiling the gut microbial composition using 16S illumina sequencing on a HiSeq 2000. We're just starting to sequence using a MiSeq using primers from Jeremiah Faith (an investigator at Mt Sinai). You could request the primers/protocol directly from him. The analysis is a little tricker as Qiime (gold standard for microbial 16S data sequencing program) is setup for mainly 454 pyrosequencing, but I think your bioinformatics people may be able to use it for miSeq sequencing analysis if they search the Qiime forum for help on how to do that.

Hey thanks Ernesto! Also, are you also at UNC? Maybe I can actually meet you in person at some point.

Our lab uses the primers designed by Caparaso et al but we had to modify them a bit to make them more robust on a Miseq system. We put the Miseq adapters at the end, next barcodes, followed by 12N's and then the 16s primer. We found that the 12N's were necessary (particularly in the V1-V2 region of the 16s gene compared to the V4 or V3-V4) to increase the "diversity" of the sample so that we could optimize the number of clusters and reads we got from the Miseq platform. Otherwise, the 16s gene is so similar between different species it affects the cluster identification process on the Miseq/Hiseq platform resulting in lower reads. I'd also suggest spiking in a high % of phiX to the sample you are going to sequence.

As Joy mentioned, there are several protocols for preparing you library and it's important to make sure that the amount of DNA from each sample is equal - otherwise the diversity matrices used in analysis can be adversely affected. I can help with some protocols for this if Ernesto's lab hasn't provided some to you already.

As far as the analysis goes... there are many ways to do this. Many people use Qiime, but as Ernesto mentioned, it requires a little tweaking for Miseq data. However, if you're new to Linux, it may be a good option because they have excellent documentation and instructions for the command line. I don't think MG-Rast will help for 16s data is it's designed for whole genomes. We stitch together our own pipeline to generate the OTU table, which requires a bit of know-how in command line, but do our final graphical analysis in Explicet (free download here: http://www.explicet.org/, http://www.ncbi.nlm.nih.gov/pubmed/24021386) and R. The following website might also be helpful in generating an OTU table if you don't want to use QIME: http://www.mothur.org/wiki/MiSeq_SOP

Good Luck!

Hi Amanda, I think you already have enough info about the library preparation, further for analyses you can use MEGAN for both 16s/shotgun sequences. With more details of your project, I can help you with right pipeline.