What is the best way of metagenomic analysis using multiple rDNA amplicons (e.g. multiple 18S and ITS for each sample)? What methodology would you use to reliably normalize and integrate the data from individual amplicons, regarding variable amplicon length between the taxons and variable libraries sizes? Can we employ Kraken or Bracken classifiers? Thank you in advance for any advice.
This is an interesting question, since people almost exclusively use one of the extremes - either single amplicon sequencing or whole-genome sequencing. With multiple genes or markers in the analysis, this is more similar to the latter. It is desireable to use a method that calculates something like the latest/lowest common ancestor (LCA) from the multiple markers - so k-mer based Kraken and alignment based MEGAN should be able to help. They both handle matches in the context of the taxonomic tree. However the usual pipelines will most likely need to be modified to take care of things like different library sizes, or to optimize the calculations by reducing any reference data only to subsets relevant to the amplicon sets (marker panels) used in the study.
So far, I found one paper dealing with that: Article A multi-amplicon 16S rRNA sequencing and analysis method for...
Yet the algorithm used here results in selecting just one best-discriminating amplicon for each taxon, if I understand it well... Does anybody have experience with this MVRSION approach?
If you have the same reference database for different fragment, you can think about to do closed reference OTU picking steps to minimize the possible bias due to different PCR fragments, samples, DNA extraction, etc. I think you can use Kraken for taxonomic assignment. But the important step is the pre-annotation step, i.e. seq quality filtering.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Molecular Methods
- Cloning