SEQanswers (
-   Illumina/Solexa (
-   -   poly-G in barcode reads? (

luc 02-25-2013 04:49 PM

poly-G in barcode reads?
Hi All,

we have a strange problem - fortunately only with a single truseq library. Two independent Hiseq runs resulted in only about 15% usable (demultiplexable) reads from genomic DNA samples.
A whopping 35% of the 6bp barcodes were read as "GGGGGG". These and other very G-rich sequences are dominating the barcode reads. These sequences do not match the actually used barcodes (which were 12 balanced truseq barcode sequences - no bias towards certain bases per position and balanced with regard to the samples).
The quality scores for the barcode reads are poor, as you can imagine, however the forward and reverse reads are fine (and they contain the expected sequences; there is no sign of any external contamination).
The libraries were generated PCR free and we actually only have barcoded Illumina adpters in the lab (truseq style). This is the first time we ran into this. Nevertheless it would be great to know how to prevent reoccurrences.

Did anybody ever come across a similar problem or has a potential explanation?

Thanks a lot in advance!

(We had the libraries sequenced at a service).

rogerzzw 03-12-2015 10:03 AM

we had the same problem with 1st run in our lab. Do you have a idea what's going on?

Appreciate your help

luc 03-12-2015 04:35 PM

Sorry, we never figured out what went wrong. Clearly there was a problem with the library prep (two sequencing runs yielded the same type of unusable barcode data). The postdoc re-made the libraries (using the same protocol - standard genomic DNA library prep using NEB and Enzymatics enzymes; Bioo adapters) and the results where problem-free.

rogerzzw 03-12-2015 04:50 PM

Thanks, luc.

What we did is pooling several different libraries (made by two persons) together to get sequence. It turned out about 30% of reads did not match any of the barcodes used in library prep. 6bp G accounts for a fair amount of reads that did not get recognized.

It looks more difficult here because you probably do not know who made it run. Maybe we can try to align the reference genome ( one person use human sample, the other use mouse sample), to see where it belongs to. So that maybe we can know.


GenoMax 03-12-2015 05:25 PM

Have either of you talked with illumina tech support about this? What was their take?

rogerzzw 03-12-2015 06:14 PM


Originally Posted by GenoMax (Post 162189)
Have either of you talked with illumina tech support about this? What was their take?

Have not yet.

I used kit that not come from illumina.

All times are GMT -8. The time now is 05:30 PM.

Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2020, vBulletin Solutions, Inc.