Hi all,
I’m working on a PhD project focused on fungal communities in dung samples from herbivorous birds (ptarmigan and geese) and need advice on selecting fungal-specific primers. The plan is to use Illumina sequencing (2×250 PE) for fungal diversity analysis, but I’ve encountered an issue with high plant DNA contamination.
In the past, I used the ITS86F + ITS4 primer pair (Op De Beeck et al., 2014), which worked great for root samples, but in downstream analyses we discovered that ~60% of sequences were host plant DNA.
Since I would like to avoid wasting half of the sequencing depth on plant reads (and we expect high plant DNA content in dung samples), the challenge is to minimize plant contamination while maximizing fungal diversity recovery.
- fITS7 + ITS4 is an option I’ve considered, but it excludes Mucorales (Ihrmark et al 2012), which are important in dung-associated fungal communities.
- Nested PCR is another option that my PI suggested (adding an extra initial PCR step with full ITS region ITS1+ITS4), but I’m hesitant because of the potential PCR bias and added complexity.
Are there any fungal-specific primers you’d recommend that exclude plants but still amplify a broad range of fungal taxa, with optimal amplicon sizes for illumina seq?
For anyone who’s faced a similar issue with high plant contamination how did you deal with this problem?
I’ve read the recommendations on the UNITE website and found their primer lists to be helpful, but I havent found a definitive solution that fits my needs. I also came across this post (choice of fungal primers) which touches on this issue.
Any suggestions or insights would be greatly appreciated!
1 Like
Hi @Daniel_Angulo ,
My group uses ITS to look at fungal communities associated with plant tissues (grapevines), dung of non-herbivorous mammals (Homo sapiens) and herbivorous mammals, among other things. So I can share a bit of my experience here.
Amen. Fungi-specific primers usually sacrifice coverage of fungal groups in the quest to avoid plant amplification. I would be curious to see if someone has designed a newer primer set to overcome this issue, but the older classical fungi-specific ITS primers really suffer.
This does not really offer a full solution, because the plant DNA would still be co-amplified in the first PCR, and then the nested PCR step would still bias against your Mucorales. This is my least favorite option due to the complexity, but that's just personal taste.
Another option might be to use blocking primers or PNA clamps, which block amplification of plant DNA by universal primers. These have been used quite a bit for preventing amplification of plant chloroplast DNA with 16S rRNA gene primers:
https://doi.org/10.1038/nmeth.2634
https://doi.org/10.1094/PHYTOFR-11-24-0124-SC
So maybe the approach could be adapted for ITS? But I just have to add, it can be really challenging to set up a working protocol, this approach can be difficult to work with in my experience, so we only use this where we expect significant plant contamination (e.g., solid plant tissues).
Another option, not super palatable but workable in my experience, is just to increase the sequencing depth per sample to compensate for plant DNA contamination. Losing ~50% of reads to plant DNA still leaves a lot of material to work with, and Illumina sequencing costs (e.g., with NovaSeq/NextSeq) have declined to the extent that overestimating sequencing coverage costs less than the extra time and consumables needed for some alternatives. This is what my group is doing for amplifying fungi directly from plant tissues and we are usually only seeing maybe 10-20% plant reads in the ITS data. So I would expect that in dung samples of herivorous birds the results could only improve unless if fungal biomass in the dung is low; I would not use root or other plant tissue samples as a reference point. So I would recommend just giving it a shot with a small number of samples to see what proportion of reads are from plants and host DNA vs. fungi, and then decide from there before adopting a more complicated protocol.
3 Likes
I've done this too in other biological contexts. 
The NextSeq 550 reagent costs at $50 to $60 per Gigabase of reads apparently, so it's cheap compare to getting/making the samples. I'm hesitant to 'throw money at the problem,' especially because I don't know how it works for your plant and fungal reads.
Anecdotally, I've seen people worry more about a PCR step than a bioinformatic step.
Maybe because each PCR step consumes sample and a software filter does not?
Maybe because 'added complexity' is lower for a postprocessing filter?