Great 16S Run!

[Anthony.287]

Hello all!
I came in this morning to see that the shotgun processing of the latest 16S run had finished, and wow!
72% passing filters, 1,532,128 total reads, and 627,095,720 total bases! Keep in mind this is with the FLX, not the Plus, and the Broad primers that seemed to give a lot of people trouble for a while. And, this isn't a fluke. The last 4 or 5 runs have been pretty similar, but this is the best so far.
Not wanting to brag (ok, maybe a little! ) but I thought I'd share some good news!
Have a great day!!

[JKistler]

Hi Anthony,

Well done on that! May I ask, what method you used to purify your 16S amplicons/eliminate primer dimers? Also, how many copies per bead did you aim for & what % enrichment did you get back? I ask as we've been trying to optimise to increase our % filter passes which remain quite low.

Thanks for sharing

James

[Anthony.287]

Details -
AMPureXP purification of ~90 amplicons, with the same ratios as the Amplicon Library Prep Manual.
PicoGreen Assay, in 384-well plate, of purified amplicons.
Diluted and pooled to 1E9 molecules/ul based on PicoGreen.
Purified with calibrated AMPureXP to remove fragments below 300bp (125.5ul AMPureXP:193.2ul DNA. I also did one extra EtOH wash during this purification.)
Run on Bioanalyzer high sensitivity chip.
Final concentration determined with Kapa Biosystems qPCR kit.
1E5 molecules/ul dilution made, based on Kapa results, and 0.08cpb used in a MV kit (4 X MVE).
5.2 and 6.8% enrichment, all of the beads loaded (1.4 and 1.8E6).

[MissDNA]

What kind of data do you get when using amplicon pipeline for analysis?
I am not a bioinformatician and understand very little about processing but our bioinoformatician would always use amplicon pipeline for 16S analysis because according to him if SG pipeline was used there would be too many artefacts left...I really don´t know the specifics on that.

We don´t have problems when sequencing using XLR70, however the same cannot be said using XL+. Still working on troubleshooting.

[Anthony.287]

What kind of data do you get when using amplicon pipeline for analysis?

For this run, I'm not sure yet. The amplicon data processing is still happening. For the run prior to this one, it was 50% passing filter, 900K reads, and 460M bases with the amplicon pipeline, versus 68% passing filter, 1.2M reads, and 522M bases with the shotgun.

[MissDNA]

900 k reads is a pretty good result for amplicon processing, which is expected to give up 30% less data than SG. Which data set are you going to work with?

[Anthony.287]

I've been providing the customers with both sets and letting them decide!

[MissDNA]

Yeah, we do the same. One of our clients really wanted SG processing but in the end when he started working with the data, he realized that eventhough there was a lot more reads with SG processing, most of it was garbage.

[ngseq21]

Hello Anthony,
I was interested in your approach to cleaning up your amplicon library and just wanted to ask why you did both the Bioanalyzer and the qPCR techniques. Wouldn't one of the two have been enough to quantify the pooled volume?
Thanks!

[Anthony.287]

ngseq21,
I use both the Bioanalyzer and the qPCR for a number of reasons. Due to earlier recommendations, I do a 2nd round of AMPure after the amplicons are pooled, and use the qPCR to check the concentration after that. The qPCR is great for this because it only quantifies DNA that has the correct primer sequences, instead of all of the DNA in the sample. The Bioanalyzer is basically just for checking the size of the amplicons, as well as to check for small fragments. The Bioanalyzer is used for qualifying, not quantifying, samples. Ours just hasn't been consistent enough to use for anything but checking sample sizes.
Hope this helps!
Anthony

[ngseq21]

Thanks for the info, Anthony!

[Dynamac]

Anthony,

How did you arrive at the 0.08 cpb ratio? Titration? Roche told me once that 2 cpb is good for amplicons.

Thanks!

Barry

[Anthony.287]

Anthony,

How did you arrive at the 0.08 cpb ratio? Titration? Roche told me once that 2 cpb is good for amplicons.

Thanks!

Barry

Barry,
This has been an ongoing project for about a year, and we started with the recommended titration range. That yielded a high enrichment percentage, so we started lowering the cpb ratios. I found a link to a few papers on here that used Poisson statistics and qPCR to determine the optimum cpb, without titrations. Their findings suggested that 0.08 cpb, when concentration is determined with qPCR, gave fairly consistent results. As we kept lowering the cpb, usually doing 1/2-sized titrations (2 cpb ratios per library, instead of 4) in 4-region PTPs, we started to narrow down the range. In January, we had half of a 16-region PTP that wasn't claimed, so we did an extended range titration, from 0.5 all the way down to 0.00025 cpb, and discovered that 0.05 and 0.1 cpb yielded 8% enrichment and good sequence. After that, I went ahead and used 0.08 cpb on the next few runs and they worked really well.
One thing to know is that the same cpb ratio didn't seem to work the same in all of the kits (SV, MV, and LV). When I used a ratio in an LV kit, after it had worked well in SV and LV, it performed quite poorly, although that very well could have been my fault!
Anthony

[Dynamac]

Anthony,

I found that Poisson paper. Interesting. I'm going to try the 1/12 ratio in my next SV amplicon emPCR; I'll publish the results here.

Barry

[Dynamac]

As a follow up, the PI decided to not try the 0.083 cpb ratio, and we went with 2 cpb instead. I don't know if the ratio is the reason or not, but we always get a large number of reads filtered out by the numFilteredTooShortQuality filter (about 60% are filtered out). The enrichments are a little high (20% or a little above). But the thing is, in one of the two control bead sets, the same thing happens. The CATG key control beads also lose 60% to the filter; the ATGC key control beads are fine. Not sure how to interpret this.