384 multiplexed 16S libraries - low quality

[NewbieNGS]

Hello Everyone,

I have done two 2*300bp runs on the Miseq with 384 multiplexed 16S libraries (derived from mice feces). For each run the quality of the reads has been really poor. I am even using 10% PhiX double what illumina recommends. For the first run I had a cluster density of 1109k/mm2 with 89% cluster passing filter and Q30 77% and this last run 1118k/mm2 passing filter of 87.7% and QC30 67%.

Have attached the results of the last run….Any idea what I can do have better runs? Or why the quality is so bad.

(I think it has to do with the mice that are producing my sample material live in a control and they should all have the same intestinal flora. )

[microgirl123]

This looks like pretty good data for a low diversity run, which this is even though you have a 10% phiX spike. Your cluster density is is a bit high for a low diversity run though - I'd try to drop it down to ~800-900K.

[GenoMax]

Is this a V3 or V2 run?

Like microgirl123 said this is a classic signature of low nucleotide diversity samples.

In spite of the poor quality values you may find the sequence adequate (if you are aligning to a reference). If you are looking to do de novo assembly then you would want to re-run at lower concentration.

[kmcarr]

You data look very typical for V3 PE300 runs of 16S amplicons, which is exactly why we advise researcher against using MiSeq V3 PE300 runs for amplicon sequencing. The V2 PE250 chemistry works very well for 16S sequencing but all the V3 PE300 runs we have attempted look very much like yours. We get just as much quality data from a V2 PE250 as we do from a V3 PE300 at lower cost.

[JBKri]

You data look very typical for V3 PE300 runs of 16S amplicons, which is exactly why we advise researcher against using MiSeq V3 PE300 runs for amplicon sequencing. The V2 PE250 chemistry works very well for 16S sequencing but all the V3 PE300 runs we have attempted look very much like yours. We get just as much quality data from a V2 PE250 as we do from a V3 PE300 at lower cost.

Do you know why the v3PE300 runs have problems with amplicon sequencing? In theory the quality should be better...
Fadrosh et al (http://www.microbiomejournal.com/content/2/1/6) has figures showing better quality from V3 PE300 cycles than V2 PE250 cycles (in supplementary figures).
Jon

[kmcarr]

Do you know why the v3PE300 runs have problems with amplicon sequencing? In theory the quality should be better...
Fadrosh et al (http://www.microbiomejournal.com/content/2/1/6) has figures showing better quality from V3 PE300 cycles than V2 PE250 cycles (in supplementary figures).
Jon

Jon,

That paper uses in-line barcodes (i.e. that are the first part of each read instead of contained in independent index reads) plus 0-7bp "heterogeneity" spacers. In other words they designed their library construction protocol to create as much base diversity as possible, especially at the beginning of the reads. This is different from the case I was describing in which all clusters have fragments starting at exactly the same spot of the 16S and thus have virtually no base diversity.

[JBKri]

Jon,

That paper uses in-line barcodes (i.e. that are the first part of each read instead of contained in independent index reads) plus 0-7bp "heterogeneity" spacers. In other words they designed their library construction protocol to create as much base diversity as possible, especially at the beginning of the reads. This is different from the case I was describing in which all clusters have fragments starting at exactly the same spot of the 16S and thus have virtually no base diversity.

So, are you saying the V3 reagents seem more sensitive to low diversity than the V2 reagents?

[microgirl123]

My experience is that it has nothing to do with the reagents. It's just that the extra 50 bp per read in a 600-cycle V3 kit are junk with low diversity samples (they're not great with a balanced genome either). You can get more clusters with a V3 kit because it reads more of the flow cell, which is the reason some of our customers like to use it.

So... we perform 600-cycle V3 low-diversity runs all the time, but warn our customers to have plenty of overlap so they don't need data from the last 50-bp of each read.