PCR Positive Control Error in Soil analysis of 16S

Hi everyone,

I’m running and analysis of environmental DNA of soil samples from a wetland-like ecosystem in South America (it’s called páramo, if you want to see pictures) using the V3 region of the 16S rRNA, in order to see Archaeal communities. This analysis was done at different depths of a soil core of 150 cm in depth, and there were 6 replicates for each sample to be amplified and sequenced. In each PCR “batch” I included a positive control to test everything was working in the reaction; this positive control was DNA from a Haloferax chudinovii. Before sequencing all the soil samples, I sequenced only this Archaea to confirm its identity and that the primers were working well and, as expected, it turned out to be OK. Then I sent my samples alongside the PCR controls (positive and negative) for sequencing in an Illumina MiSeq PE 150. After demultiplexing I followed the moving Pictures Tutorial for quality filtering and Taxonomic Assignment, but the positive control is not even close to being an Archaea, in fact, the taxonomy shows that it has Polaromonas sp. For the classification I have trained my own naive-bayes classifier on SILVA 132 and SILVA 138 and have also used classify-consensus-blast on NCBI refSeqs using the primers for V3 region I used, but I get the same results. In addition, I couldn’t find any Archaea in the samples either.

I feel that having this result on my positive control invalidates the results in the other samples
(regardless of presence/absence of Archaea), since it is completely different from what I was expecting, but I want to know if someone has encountered this issue in their analysis and what they have done to overcome this. I imagine that the best possible solution is sequencing all the samples again, but doing everything all over again is not feasible at the moment. Maybe somebody has a better opinion about how to approach this problem, either in the lab work processing or the bioinformatic analysis.

Thanks again and sorry for the long post.

Hello again @ancazugo,

Sounds nice!

Did you use extract-reads to trim to your primer region? I recommend searching the taxonomy file to make sure that Haloferax chudinovii is present. Even if the primer amplify the positive control, if they are a poor match for Archaea then you may need to adjust your mismatch tolerance to allow in silico "amplification" (with extract-reads)

hm... new method, new database, same bad result. It's concerning, but still does not rule out a technical error. Same possibility as before... using extract-reads may be causing Archaea sequences to drop if those primers are a poor match. You should check out the taxonomy (after extracting reads) to confirm whether or not your target sequence is actually in the database.

As a last-ditch effort, I recommend grabbing the most abundant sequences in your positive control and use NCBI BLAST to see what the top hits are. If they are not Haloferax chudinovii or remotely similar, then I worry about contamination issues. Otherwise, I would hold out hope that it's a technical error.

Let us know what you find!

1 Like

Thanks again for your advice!!!

I used extract-reads on every database with the primers I used for amplification of the V3 region (U341f: CCTACGGGNGGCWGCAG; 534r: GWATTACCGCGGCKGCTG), and I checked after extracting reads and several species of Haloferax are in the ref-seq file.

In addition, I manually blasted the sequence from the positive control and it still resulted in Polaromonas.
I was hoping I could find something by changing some parameters in QIIME2, but I guess the problem is contamination in the samples.
Probably the best thing I can do is to do the lab work again, since my results are not reliable :expressionless:

Thank you very much for your help!!!

Sorry to hear that @ancazugo! Sounds like this is almost certainly contamination. It may be reagent contamination... I'd recommend doing some sleuthing work, maybe also discuss with whomever performed the sequencing to figure out where Polaromonas might be creeping in.