truncated adapter sequences in cloned libraries

[greigite]

Moving this up to a new thread for more visibility... . We cloned and Sanger sequenced some standard genomic PE libraries and are seeing truncation of the adapters from the 5' end in some libraries, as well as a larger number of molecules containing full PE1 sequences than full PE2 sequences. We used NEB library preparation kits and some custom barcoded adapters from Invitrogen, which contained the phosphothiorate modification and were purified using desalting. Following triplicate 4 cycle PCR (we also tested 8 and 12 cycle PCR but saw little difference in this respect), we cloned several of the libraries into the Fermentas blunt end pJet vector and Sanger sequenced several transformants. While some clones looked normal, several had truncations off the 5' end of PE1 (see attachments). These truncated sequences did contain some bases from the PCR primer on both sides, meaning that we weren't just sequencing unamplified ligated molecules, and there were no adapter-only clones. I haven't been able to come up with a reason why the primers themselves should be truncated aside from being poorly synthesized- but that surprises me since I have not had similar problems with Invitrogen oligos in the past. We are not sure what steps to take next- should we just qPCR quantify the libraries with Kapa so we know how many correct molecules are present and use that to estimate loading amounts? Or did we go wrong elsewhere in library prep?