Barcoded PE adapters for multiplexing up to 12 samples

[Pepe]

Hi all,

We would like to ask your opinion on this issue. I'm sure that people is more used to thinking about this than we are.
We are designing our own barcodes so they are as short as possible. We made them so that they have enough redundancy to easily discard any non barcoded sequence, and so that no mutation will make one barcode transform into another.
The result is a set of 12 3-bases barcodes (plus the extra T necessary for the ligation, so 4 bases total). We are pretty confident that that's what we want to use, but, for what we have seen, no one is using barcodes this short.

Can anyone come out with a reason for not using these barcodes?
Any help is very much appreciated.

Thanks in advance.

-Pepe

[nilshomer]

Hi all,

We would like to ask your opinion on this issue. I'm sure that people is more used to thinking about this than we are.
We are designing our own barcodes so they are as short as possible. We made them so that they have enough redundancy to easily discard any non barcoded sequence, and so that no mutation will make one barcode transform into another.
The result is a set of 12 3-bases barcodes (plus the extra T necessary for the ligation, so 4 bases total). We are pretty confident that that's what we want to use, but, for what we have seen, no one is using barcodes this short.

Can anyone come out with a reason for not using these barcodes?
Any help is very much appreciated.

Thanks in advance.

-Pepe

We have used barcodes from 4-8bp in length with no problems on Illumina.

[upenn_ngs]

I am working on 6bp bar-coded PE prep followed by pooled capture with Agilent SureSelect and solexa sequencing. I would like to know during custom oligo design, what *minimum* modifications are necessary for asymmetric product?

[simonandrews]

The result is a set of 12 3-bases barcodes (plus the extra T necessary for the ligation, so 4 bases total). We are pretty confident that that's what we want to use, but, for what we have seen, no one is using barcodes this short.

We've never run as many as 12 samples in a lane, but we've run 4bp barcodes.

The only problem we've had has been due to the loss of sequence diversity in the first 4 bases giving us really high purity filter losses. This is actually less of a problem the more barcodes you use. There was a separate thread about this in the general sections of the site.

http://seqanswers.com/forums/showthread.php?t=1460

[Pepe]

Thanks Simon,

We are now running 8 samples per lane and we don't see many more purity filter losses compared to when we did not use barcodes.
I have been a very interested observer in the the threat that you suggest. Regarding that, I am worried about the 4th base being the same in all our barcodes, therefore having a higher chance of the pipeline not being able to discern between close clusters, as you explained in the thread. Again, we are roughly getting the same number of 'purity filtered' sequences in each run, so maybe I don't need to worry much about it.

A couple of questions:
Are your 4bp barcodes 3 bases+a check (AAAT, CCCT, ATGT, TGTT, GCAT...) or 4 bases all different among barcodes (ATGC, ACTG, AAAA...)?
I can't think of anyway of testing if the use of these barcodes is reducing the amount of sequences that we get. Do you know of any?

Thanks for your reply.

Pepe


We've never run as many as 12 samples in a lane, but we've run 4bp barcodes.

The only problem we've had has been due to the loss of sequence diversity in the first 4 bases giving us really high purity filter losses. This is actually less of a problem the more barcodes you use. There was a separate thread about this in the general sections of the site.

http://seqanswers.com/forums/showthread.php?t=1460

[graveley]

We've done quite a bit of barcoding. Using a 3 nt barcode, we've actually multiplexed up to 64 samples in a single lane. We have started using 6 nt barcodes which are composed of 2 repeats of a 3 nt barcode. This helps to reduce mis-assigning reads to a sample due to a single sequence error. This has all been done using custom primers where the barcode is the first sequence read, then the rest is the actual sequence of interest. We've also done multiplexing using the index read method and done up to 12 codes and this works great!

[micro]

What's the annealing program that you use for the custom adaptors? I'm wondering if you have to purify annealed adaptors before the ligation or you just use equal molar adaptor, anneal and directly use in the ligation. Thanks!

We've done quite a bit of barcoding. Using a 3 nt barcode, we've actually multiplexed up to 64 samples in a single lane. We have started using 6 nt barcodes which are composed of 2 repeats of a 3 nt barcode. This helps to reduce mis-assigning reads to a sample due to a single sequence error. This has all been done using custom primers where the barcode is the first sequence read, then the rest is the actual sequence of interest. We've also done multiplexing using the index read method and done up to 12 codes and this works great!

[Pepe]

Here's what the person that did this replied:

Basically the later method:

-I mixed the two primers, equal molar amount, in annealing buffer (10 x anneling buffer: 100 mM Tris-HCl pH 7.5, 10 mM EDTA, 500 mM NaCl).
-Incubate at 95C for 2.5 min, then cool it down to 25C (- 1C/30 sec), then at 25C 10min, then 4 C.
-Quantification
-Dilute the products by adding water to adjust the concentration to 360 ng/ul, because that's concentration of the adapters in the Illumina kits.

We can not exclude the possibility of contamination of un-anneled primers, but this is not usually a big problem.

[micro]

Thanks, Pepe. I'll give it a try.

[seqgirl77]

Hi, I have a question about the extra T base in barcodes. I see why the extra T on the 3' end of the PE2 adapter is necessary (for ligation), but people seem to also have a T after their barcodes. Why is this?

Thanks!