FelixKrueger
commented
3 years ago Hi Matt,
I think the short answer is: Trim Galore isn't explicitly aware of this new PCR-free kit, and if it did the right thing there chances are it was by pure luck. If the Nextera sequence is detected, the same sequence is being used for both R1 and R2, here CTGTCTCTTATA.
My gut feeling tells me that your R1 sequence is a perfect match to the Nextera sequence, so that is fine:
CTGTCTCTTATA
CTGTCTCTTATACACATCTYour R2 sequence
CTGTCTCTTATA (Nextera)
ATGTGTATAAGAGACA (adapter sequence you specified)
.TGTCTCTTATACACAT (reverse complement of adapter sequence you specified)This would mean that the R2 adapter matches the Nextera sequence in 11 out of 12 positions, and if any of these last 11 positions were found at the 3' end of reads, it would trim the correct sequence.
If the sequence was found earlier in the read (not right at the very end), then the first C in the Nextera adapter sequence would also have to match for the adapter to be trimmed as contamination as a full match. There are now options:
1) If we trust that the sequence you gave above is correct and complete, the C used by specifying the Nextera adapter sequence (which occurred automatically) is not from the adapter, that C is matched to a genomic base at that position. If we assume that C occurs in ~20% of cases you would get a 'perfect adapter match' in ~20% of (c|b)ases, but in 80% of cases there would be a mismatch (T, A or G). As the default error profile of Trim Galore is 0.1, it would allow 10% of bases in the adapter sequence to mismatch, which in the case of a 12bp adapter sequence, such as the Nextera sequence, would mean 1 mismatch is tolerated. In other words: Even though the sequence used isn't exactly correct, it just luckily does the right thing anyway.
2) Is there a possibility that the sequence you specified above is short by one base pair? If that were true, and that base would be a G, then it would be a perfect reverse complement of the Nextera adapter sequence:
CTGTCTCTTATACACATCT+ATGTGTATAAGAGACA GWhat also often happens is that the adapter sequences are given relative to the top strand of the DNA, which isn't necessarily the way sequences get read when they come off the sequencer.
If we would further assume that the sequence you gave are the R1 + R2 adapters, but in fact the R1 adapter occurs at the 3' end of R1, and the R2 adapter at the 5' end (which would match up with the illustration on slide 4 of Illumina's tagmentation, we could end up with a situation that would like this (reference to the top strand, I means insert base for top strand, and X is the reverse complement to I on the bottom strand):
5' ATGTGTATAAGAGACA G IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII CTGTCTCTTATACACATCT 3'
3' TACACATATTCTCTGT C XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX GACAGAGAATATGTGTAGA 5' This would perfectly fit the following model:
- one read would start reading the insert bases
IIIIIIIIIIIII..., and then reads into the Nextera adapter on the 3' end. This is what is auto-detected, and used to trim both R1 and R2. - The other read also reads insert bases, here
Xin 5'->3' direction, and also reads into the a sequence that looks like the Nextera sequence.
I realise that there are quite a few if's in this assumption, but it would make so much sense... Even if the additional G that I invented to fit my flow of words better was not really present, it would most likely still do the right thing as outlined in option 1) (maybe this is even true for any Nextera based tagmentation trimming)?
I hope this is more helpful than it is confusing, do let me know if you've got any questions.
andrews-matt
I have used Trim Galore v0.6.5 to trim the adapter sequences from some paired-end WGS data that was prepared using Illumina's DNA PCR-Free prep, which uses adapter sequences CTGTCTCTTATACACATCT+ATGTGTATAAGAGACA. Although your documentation does not specify the second sequence, looking at the clipped reads it does appear that Trim Galore was able to automatically detect and trim the reverse complement (TGTCTCTTATACACAT) from the reads in the R2 file.
I understand that the first sequence was detected from the 12bp as being identical to the Nextera adapters so that makes sense, but I seek to understand how Trim Galore was able to detect the reverse complement of the second sequence considering that Illumina only released the PCR-Free prep in the past 12 months and I am using an older version of Trim Galore. Mainly I want to understand whether this is expected behavior or whether I should manually specify the adapter sequence just to be safe.
Thank you.