Ion Torrent adapter artefacts

[lexx_04]

Hi guys!

We recently began working with Ion Torrent PGM. At first we were thinking that server cuts off all adapters and barcodes and we are getting our clear and clean data out. Then we began noticing that sequences do have something in common at the end.

Through the process of googling and brainsturming, we found that about 40% of all our output data has got REVERSE P1 + B adapters at the end of our sequences.

P1: ATCACCGACTGCCCATAGAGAGG
B: CTGAGACTGCCAAGGCACACAGGGGATAGG

This is kinda wierd and we are not sure how to deal with that. We are thinking that this might have a negatine influence on our results increasing “unknown” and “unclassified” domens.

Also, quite mysterious that we see a lot of REVERSE primers but very little FORWARD ones.

Has anyone faced similar problems? What do you do about it?

Will be very thankfull for any ideas/ thoughts/ advices!
Kind regards,

Alisa

[Markiyan]

First have a look at your library prep. Than on processing.

First it would be help full if you could provide a bit more info about the sample/library prep used.

Actually adapter artefacts can come in many variations, esp if you have a bit of nuclease activity during the library prep steps before/during the ligation steps (adapters may loose some of the 3'/5' nucleotides, form chimeras, ect.

When they get sequenced, the instrument's data processing software is likely to miss them during processing, and they get into the downstream analysis...

Some analysis (like mapping to a reference) would tolerate them, but other ones (de novo assembly) may be very sensitive to them.

So convert the data to fastq format and run it through the fastqc (up a kmer size to 9 or 10 -k 9) and have a look on overrepresented sequences.

Than grep out some reads with those sequence and and have a more detailed look.
Use grep -c to get the number of matches

Once you know your full contaminant sequence - remove it with cutadapt or trimmomatic.