[arrchi]

Thanks again. I think these sequences (CGTCTG and TATCTCGTATG) will be removed if i do adapter trimming?

Do you have experience of using any software to trim adapter sequence?

[moritzhess]

In my opinion the most convenient and sensitive but also potentially slowest way is to align the illumina adapters using e.g. SSAHA2 which will even detect adapters when there is a rather high error rate in the data. The SSAHA2 output can then be parsed to cut the first base in a read with aligned adapter or to set the PHRED Score to zero from this position on.

UPDATE:

In fact it is not that slow: 2 min per 1M reads.