structure of an adapter-dimer

[woshiliangliang]

When preparing an Illumina library, we always talk about "Adapter-dimers", but what is the Specific structure of a dimer? And how can I detect or observe it?
Any ideas?

[Brian Bushnell]

For artifact-dimers, if the sequences are identical and you have paired reads, you can find them by merging reads via overlap and looking in the insert size histogram for sharp peaks at specific insert sizes (e.g. a sharp peak at 62bp would indicate dimers of some 31bp sequence). These reads should not map to the reference.

Generally, if you have sharp peaks in the insert size histogram from overlap that are not present in the insert size histogram from mapping, these are probably artifacts.

If you know the artifact sequence, you can also find (or trim, or filter) artifacts by searching for reads containing the sequence in question using a tool like BBDuk. An adapter dimer should have adapter sequence starting at position 0.

[woshiliangliang]

thx Brian,
but i just want to know what is the specific structure of the artifact-dimers,
i couldn't understand how two adapters ligate with each other.
could you simply describe the molecule mechanism?
Thank you.

[Brian Bushnell]

I'm a programmer with some knowledge of chemistry. Everything I know about sequencing has been taught to me by people more familiar with biology and/or kinetics. So no, I can't tell you anything about which non-genomic artifacts will preferentially ligate; sorry! All I can tell you is how to detect them. But I'm sure someone on this forum can better help you.