Forward adapter at end of forward paired end read?

[darthsequencer]

So I was given a metatranscriptome prepped with the Nugen Ovation RNAseq kit.

The metatranscriptome was sequenced on a HiSeq 2000 paired end.

In the first read a lot of the sequences ~20% have the forward adapter at the end of the read - this is it's sequence:
GCTACGGGATGACTTGTGATTAGGGGTGAGAATGTCGGCCTGAGAGATCGGAAGAGCGGTTCAGCAGGAATGCCGAGACCGATCTCGTATGCCGTCTTCTGC

I'm just wondering what could cause this -
Here's an example read with the adapter in it. If anything shouldn't it be at the front of the read? Why would it be at the end and why is there still more sequence after it?

GCTACGGGATGACTTGTGATTAGGGGTGAGAATGTCGGCCTGAGAGATCGGAAGAGCGGTTCAGCAGGAATGCCGAGACCGATCTCGTATGCCGTCTTCTGC

Thanks

[BAMseek]

Reading into the adapter sequence can happen when the fragment size is less than the read length. This would show up as a partial adapter sequence at the 3' end (right side) of a read, as you have observed. There are quite a few tools out there that will trim the adapter portion of a read (see here for a discussion, and also search for threads here on seqanswers). I'm not sure how well these tools deal with reads that go through the entire adapter sequence and beyond - one possibility might be to pad the 3' end of the adapter sequence with Ns before running one of these tools.