How to distinguish between 16S chloroplast reads and real bacteria ones?

Question

[Raito92]

Hello,

In an old experiment dating back to 2014, some 16S libraries from plant samples were prepared, with the aim to identify the microbial community living/contaminating the surface of the fruits.

The primer choice was the couple 27F and 533R.

Sadly, once analyzed with QIIME2 and clustered, out of 12 samples, 10 had chloroplast DNA contaminations percentages up to 99% (80-99%), making them completely useless.

I'd like to run the experiment again, with Illumina technology, addressing the problem of the primers not being specific enough to distinguish between plastid and bacteria DNA.

At the moment, the only scientific paper directly addressing the issue I was able to find was this one, but it's quite an old one.

Any suggestions here? Thanks in advance.

[Mark]

Just to confirm, you want to resequencing the sample specimens but with illumina 16S sequencing and you want to know of a way to differentiate between bacterial and chloroplast 16S?

QIIME2 has a filter module you should checkout: https://docs.qiime2.org/2018.11/tutorials/filtering/

You can also filter with phyloseq https://joey711.github.io/phyloseq/

It might be possible to align all illumina reads to chloroplast 16S, then use any reads that don't align in downstream analysis.

[Raito92]

Thanks for the suggestion, but I wasn't referring to this. I have already tried the qiime taxa filter-table function and it works like wonder, but it just removes the reads I'm not interested on from the QIIME 2tables.

The problem isn't about filtering data, but about obtaining data with a lower contamination rate since the very beginning. So my issue is about primers choice.

[Mark]

Right, apologies I misunderstood. Hopefully someone can provide some help.

Answer 1

[andres.firrincieli]

Hi Light,

we used the method described here to lower the 16S contamination from chloroplast and mithocondria

[Raito92]

Hi, thanks for your advice! I checked your paper out, but it appears not to be available for free, and the abstract just says very little. Could you expand on such methods, even through a private message? It would be greatly appreciated! Thanks in advance!

[Mark]

try this link http://sci-hub.tw/10.1038/nmeth.2634

[andres.firrincieli]

Hi, Unfortunately our samples were processed by the Joint Genome Institute (JGI), so I know very little about the library preparation methodology. I am the guy who processed the raw data. If you give me some time I can tell you how many reads per samples were assigned to chloroplast and mitochondria.

[Raito92]

This would be greatly appreciated, thanks! However, I'm mainly interested in the primers choice.

[andres.firrincieli]

Hi light

The number of reads assigned to plastid rRNA ranged from 10 to 30%. I guess these values are highly dependent by the nature of samples and by the extraction protocol. Therefore, I would test a couple of samples to check if the methodology is working properly