SeqrunAdapter contamination and when to trim
Adapters are the synthetic sequences ligated onto every library fragment. When an insert is shorter than the read length, the sequencer reads off the end of the insert and into the adapter — and those bases need removing.
Why adapters end up in reads
A read has a fixed length — say 150 bp. If the insert between the adapters is longer than that, fine: the read stays inside the insert. But if the insert is only, say, 120 bp, the last 30 bases of the read spill into the 3′ adapter. Short inserts (degraded input, small RNAs, over-fragmented libraries) are the usual cause, which is why adapter content rises as insert size falls.
How it shows up
A QC report flags it as an over-represented sequence or a rising adapter-content curve toward the 3′ end of the reads. The over-represented sequence is often the start of a known adapter. MatchingTTTTAGATCGGAAGAGCACACGTCTGAAC against the common set, for instance, finds AGATCGGAAGAGC — the Illumina TruSeq/universal adapter — sitting inside it. CTGTCTCTTATACACATCT is the Nextera (Tn5) transposase adapter.
When to trim — and when it matters
Trim when adapter bases are actually present; if inserts are comfortably longer than the reads, there may be little to remove. Adapter bases left in reads are not real genomic sequence, so they cause soft-clipping or mismatches at the read ends, hurt mapping, and can create spurious variants — the effect is worst for small-RNA and amplicon work where inserts are short by design. Identify which adapter you have, then trim that specific sequence with your tool of choice. Note that real reads carry sequencing errors, so trimmers match with some mismatch tolerance rather than demanding an exact hit.