problems with cloned library molecules

[sherisim]

We are finishing up a batch of genomic libraries for PE sequencing
prepared using NEB's kit and some barcoded adapters from Invitrogen,
and have amplified them in triplicate with 4 cycles of PCR (we also
tested 8 and 12 cycles). The adapters and primers are the standard
Illumina sequences with a 6 bp barcode attached. We cloned and Sanger
sequenced a few transformants to check the amplification and
barcoding, and it looks like many of these library molecules show odd
truncations on the 5' end of one or both adapters. The sequence
adjacent to the genomic fragment is fine. I am not sure what could be
causing this- it almost looks like the ends are getting chewed up
during PCR or the cloning itself as all transformants contain at least
some of the sequence added on in the PCR step (i.e. they are not just
the ligated adapters). Can anyone chime in on what you have seen in
library clones in the past- does this sound at all familiar? We are a
bit mystified at the moment. It seems to me that qPCR using the PCR
primers should at least accurately quantify the total number of
complete molecules in the library, so we should be able to adjust
cluster densities accordingly, but it is worrying and will negatively
affect our ability to accurately pool samples.

[elaneyk]

What kind of modifications/purifications did you get for your adapter
sequences that were synthesized by Invitrogen? I get my barcoded
oligos synthesized by IDT dna and include the modifcations that the
normal illumina adapters have (which I'm pretty sure are there to
prevent chew back). I also get them HPLC purified.

Elaine