Hi there!
I have a very broad, yet not so much technical question. I've been tampering with sequences coming from deep metagenomic study of gut microbiome: approximately 9.5Gbp/sample. I've used a standard [SOP][1] (see link) for analysis, namely to come up with:
1. Detailed taxonomic composition
2. Functional potential of microbiota - mapping sequences to a KEGG database.
It uses metaphlan2 (clade-specific marker gene for phylogeny investigation) and humann1 (although I have also computed results using humann2). For downstream analyses I've used several statistic packages to compare obtained profiles: - STAMP - LeFse - MaAsLin (to reduce the metadata that are introducing "noise", like age, bmi, ethnic group)
This study is concerned broadly with differentiating between health and in-disease state microbiota. We've got 16S rRNA gene survey also available for this whole experiment. I am thinking now about a StrainPhlan, a tool to investigate SNPs between certain microbial strains belonging to a species of interest.
I'd to know whether there might be something else I could explore, and I've missed here. Any suggestions, comments, questions are welcomed.
Best, Robert.
Edit:
I was thinking of identifying pathogens as opposed to neutral/beneficial microbiota, or specific strains of bacteria. I've read somewhere that "PATRIC" database could be used for that. Perhaps then identifying certain SNPs in strains might at least partially drive analyses further for in-disease microbiota.
https://github.com/mlangill/microbiome_helper/wiki/Metagenomic-standard-operating-procedure
That might be interesting to see, I wanted to test that, but Whole Genome Sequencing experiment comes first now.
Is MEGAN6 superior to some extent to metaphlan2 + humann1/2?
I have not used MEGAN6, personally I have been happy with the results from metaphlan and humann, another interesting tools would be kraken and centrifuge both from same group, kraken has awesome visualization via krona tools.