SureSelect adapter trimming

[swNGS]

Hi,

This is a simple question, but I cant find mention of whether or not to routinely trim adapter sequences from Agilent's SureSelect PE library prep kits.

Do any of you routinely apply an adapter trim to your SureSelect samples prior to alignment as part of an existing pipeline?

I have a well defined process for adapter trimming Haloplex FASTQs as there is usually a fair bit of adapter read-through since the amplicons are often shorter than the 2x150bp PE reads that we run.

For SureSelect I wanted to instigate a routine adapter trim as part of an automated pipeline, if for example the genomic DNA had been over-sheared, leading to smaller than ideal fragments, which would result in adapter readthrough (in the same way that I currently have to deal with read-through in Haloplex assays).

Does anyone do this for SureSelect, and if so, what are the adapter sequences that you use?

Thanks

[Bukowski]

I've never seen adapter contamination in any SureSelect sample I have processed, so I would suggest it's not necessary. With HaloPlex it is absolutely essential.

[swNGS]

Upon examination of insert metrics generated by Picard tools, I get a small 'notch' or 'step down' in numbers of insert sizes smaller that the 150 base boundary point (2x150 PE in use), with a rapid drop off below 150bp (which I would expect). I suspect that the 'notch' which I put down to an improvement in alignment of PE reads after 150 bp is down to adapter contamination in the 3' end of reads where the actual insert is less than 150bp, leading to impaired alignment. Its also possible that since these reads are getting smaller towards he lower extreme of the range of insert sizes will not align so well just because there's less sequence to align... If this was the case I would expect the drop off to be smooth. My mini 'step' at 150 bp suggests otherwise.

FASTQC does not any overwhelming over-represented sequences flagged as adapter, but I dont know if they would be pre-defined in its own database?

I realise it wont make a massive difference to the alignment, but is niggling me!
Without the adapter seq I dont know how to check... any ideas?

Thanks

[swNGS]

Several libraries later, I have encountered exactly this issue with an over-sheared SureSelect library and probable adapter read-through, which aligned very poorly (bwa_mem). Agilent won't release the adapter sequences, so I was hoping that someone out there would know what they are?

Any takers out there?

[fkrueger]

The k-mer plot of FastQC might yield a clue what the adapter sequence might be.

[Jon_Keats]

Agilent adaptors are exactly the same as the illumina truseq less the 8bp barcode in the case of XT2 adaptors replaces the 6bp barcode in truseq. Ultimately there is little room for changes as there is sequence requirement to bind the flowcell and for the sequencing primers to bind for Read1, Index1, and Read2.

[swNGS]

Okay, that's good to know.
I found this description of the TruSeq adapters here

http://genomics.med.tufts.edu/documents/protocols/TUCF_Understanding_Illumina_TruSeq_Adapters.pdf

I'm using XT kit, and not XT2? Should I be approaching this any differently given that? Is the barcode in the same position for the XT kits?

So, am I correct in thinking that to account for any read-through, I should trim the reverse order of universal adapter from the 3' end of the insert, and the reversed sequence of the adapter on the side with the index, nearest the insert from the 5' end?

Thanks for the pointer

[ECO]

The final adapter in XT (after post-cap PCR) is identical to Truseq adapters.

[jordi]

Hi all!
No adapter trimming with sureSelect and good variant calling results.
However I've found what swNGS said regarding reverse complementary adaptor sequences resulting from amplicons shorter than read length using HaloPlex. I don't know how to trim such sequences.
Ideas will be appreciated!
Thanks