Barcode Leakage?

[PersianExcurzion]

Has anyone had a problem with the SOLiD 3.0 barcodes presenting in a spot that the barcode was not used in? I should add we are using the SOLiD 4.0 now with all current chemistries, but still use the 16 barcode set from the 3.0.

We suspect either barcode leakage (one barcode being mistaken for another during sequence analysis) or the fact that our barcodes could be contaminated. I wanted to know if any one else was having this problem though.

[dsidote]

We have been having barcode issues since moving to SOLiD4. We have found BC contamination in our samples, this was verified for us by AB. We used the higher number BCs (17-96) and found contamination of 1-16 in the samples. We changed kits and the problem remains. I suspect its a manufacturing problem or a software problem.

Also, whenever we have a sample with BC3, it seems to be significantly over-represented compared to the other samples on the slide. AB claimed our kit was contaminated but we have been able to reproduce the problem in libraries made with different kits, in different labs, by different people.

[pmiguel]

How much contamination? Greater than 1%? Does the issue persist if you use setting that call a given barcode only if the sequence is perfect (zero mismatches?) Are these RNA or DNA samples?

Another possibility would be amplicon contamination from earlier libraries. Or have you never used BCs 1-16 in your lab?

--
Phillip

[PersianExcurzion]

Yes, greater than 1%. We try to detect heteroplasmy in DNA samples down to 1%, and this causes an irritating background. We can tell its the other samples that are being mixed up in the spot, and someone needs to determine if the heteroplasmy we see in a patient is real, or if it is the homoplasmy of another patient leaking in.

We recently changed the barcode stringency, and that made a significant difference. But it has not fully solved our problems. We have ordered the 96 barcode kit to see if that will make anything better. However that is on back order till late march.

It sounds like its just a more relaxed chemistry with the SOLiD 4.0. Which almost all labs would not be concerned with an up to 5% background when dealing with nuclear genes, but for us its a huge difference. Maybe that's why the jump to the 5500XL over the 4.0 upgrade.

[pmiguel]

I do not like the sound of this.

One thing is that for DNA, until recently, for oligos beyond the initial 16, we just bought our own oligos and annealed them to form adapters. Once Life Tech started offering them for sale, we started just buying theirs. But if you suspect their QC, you could purchase the oligos directly from a trusted oligo synthesis source.

But the problem with just blaming this all on Life Tech is that mixing in many cases could have occurred just as easily during the DNA preps, or at any point up to the point the bar codes are added. How carefully have you examined your pipeline? Are you always using plug-seal pipette tips, etc. to rule out sample mixing?

If you really think that the software may be at fault, the actual bar code sequence files are generated during base calling. So, those could be checked to make sure there is not some bug in the downstream bar code assignment.

Finally, the 5500XL jump seems like a play from the normal Applied Biosystem's playbook. Seems nearly identical to the situation with their first 96-capillary sequencer, the 3700. This instrument was slated for release of a new polymer, POP37, that would lengthen reads. But instead the entire instrument was discarded in favor of the 3730XL -- a completely new instrument. The play is:

(1) Start from behind. (Molecular Dynamic's had a year, at least, lead on 96-capillary instruments with their "Megabace".)
(2) Release an instrument that is basically a prototype. (The 3700 back in 1999 or so, the SOLiD v1 more recently.)
(3) Swap it out for a more mature instrument based on information gleaned from running the prototype instrument. (The 3730XL and the SOLiD5500XL).

The difference being that the 3700 had significant advantages over the megabace--it was able to beat it in nearly every aspect right out of the gate. Most of the large genome centers bought 3700s instead of Megabases. (With the JGI's predecessor being the obvious exception.) By the time the 3730XL was released Applied Biosystems was largely mopping up with Megabase sales confined to limited areas and markets outside the US. Not at all the case with Illumina -- who are largely trouncing ABI currently.

--
Phillip

[SeqAA]

Sorry Mate. I'd be 99% sure this is the library prep and not mfg issue. I've seen bracode mixups on different platforms, and it's always tracked back to user error.

And it's the same chemistry in the 5500 akaik. you will not see any different barcode assignments there.

Make sure the libraries are being made in a clean room, or at least a clean area.

[pmiguel]

So how could the purity of the bar code adaptors be checked? Limited digestion with Snake Venom Diesterase followed by mass spec? (I am not a mass spec guy. This is just a method for sequencing an oligo prep that I heard once.)

--
Phillip

[SeqAA]

heh. You could to a ToF digestion, but i really don't know the sensitivty. it's been a while.

You could design an end point pcr test.

If there is a chance of contamintion, you usually have to start from scratch. not really worth the time and effort to QC.

[pmiguel]

Yes, it would be to exonerate or bring to justice Life Technologies. Not that they accept the judicial authority of a SeqAnswers court.

They are providing the adaptors/oligos. So, if the OP doesn't trust their QC, they would have to have them synthesized themselves...

--
Phillip