I'm starting to plan a new project to characterise the biogeochemistry of a site where radioactive waste was buried more than 50 years ago: water flooding a chamber where it is allocated and sorrounding soil.
I have suggested to do proper metagenomics aside of 16S and 18S amplicon sequencing to provide a broad image of the potential metabolism occurring on the site, with especial regards to microbial elemental redox transformations, e.g. iron, and sulfur cycles. However, I read somewhere (I can't locate it now) that the presence of eukaryotes in the samples may cause problems for a good coverage.
Taking in account that oxidation/reduction of Fe, S, N species amongst others are mainly driven by prokaryotes, but also considering that the carbon cycle could be dominated by eukaryotes, should the later be filtered out? shouldn't?
I have experience with amplicon libraries to assess diversity and with whole genome sequencing but with the meta-omics techniques I'm a complete newbie.
By the way, is it worth to do rDNA amplicon sequencing if I'm going to do metagenomics?
After you do the sequencing, you can filter out the eukaryotes using software. I used MGRAST server to analysis the metagenome results.
The problem with the Eukaryota fraction is that it will "steal" a lot of the reads, seeing how eukaryotes have bigger genomes compared to prok. Thus, you'll be spending a lot of your coverage sequencing euk and might be blinded to a lot of the prok fraction.
This does however very much depend on the abundance profile of your community. It might be worth having a quick "prospective" look using 16S and 18S just to get an idea of the ratios of the two. If it does look like you have a lot of euk, then you should consider somehow cleaning those out (no idea how you actually do this, but i'm guessing a size filtering of something).
If you don't want to bother with a prospective investigation you could always just separate the two and do metagenomics on them separately. I think i would prefer this approach, but i'd have to think about exactly why.
There are many ways to filter out eukaryotes from the short reads. Most of the tools likein MGRAST have inbuilt functions to do so. You could also do quick blast and get rid of the reads if you want to use you own more stringent criteria. But, since eukaryotes have much bigger genome and have potential to effect the output of the sequencing, it should be cleaned as much as possible before sequencing the sample. Usually the sequencing facilities have some protocol for QC before sequencing. Also the company who developed the technology you are planning to use might have instructions for removing eukaryotes. This is the a problem in every type of metagenomic experiments.
Thanks, Paul. That's actually what I was asking for... How'd you do the "prospective" examination of 16S/18S? qPCR?
from the water you can use a 0.8 micron filter to 'mostly' separate prokaryotes and eukaryotes. However, do not ignore the surfaces- biofilms will be very important in oligotrophic situations such as groundwater. You might even place some plates or slides to grow biofilms to sample.
VERY GOOD RESEARCH HOWEVER I AM UNABLE SUGGEST ANYTHING MORE. gADEWAR
Indeed Eukaryotes will steal some of your reads. Now you can get good prices with very large coverage if you chose to use Miseq for example. Maybe that could work. Also, have a look at that paper from Logares et al., 2013 regarding the use of 16S-18S amplicons before metagenome studies. Finally, regarding the cleaning out of eukaryotes, if you do size filtering you'll probably loose some relevant prokaryotes too as mentioned above. I would go for the whole community with large coverage and eventually get rid of the eukaryotes on MG-RAST.
And good luck, starting in metagenomics is not easy...
https://www.researchgate.net/publication/257532416_Metagenomic_16S_rDNA_Illumina_tags_are_a_powerful_alternative_to_amplicon_sequencing_to_explore_diversity_and_structure_of_microbial_communities
Article Metagenomic 16S rDNA Illumina tags are a powerful alternativ...
Xabier, qPCR would be the easiest way to go. I'm thinking there should be a "faster" way if you have many many samples. Maybe a heavily multiplexed MiSeq run (something like 96 - plex), though i'm not really sure how that would work.
If you were to take the V3-5 primers for 16S, you should be able to assign all the resulting fragments (OTUs) to kindom level (don't actually know if this is true, but for my sanity i'll just assume it is). So then whatever doesn't get assigned by say, RDP, would be Eukaryota. Does that make sense?
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools