Next-generation sequencing library preparation is a long chain of small, precise steps, and it punishes error in a particular way: mistakes do not just add, they compound and they hide. A slightly off ratio in a bead cleanup, a normalization that used a shaky small-volume transfer, an enzyme mix added a little unevenly, and the failure shows up much later as a skewed library or a sample that drops out of the pool entirely, long after the offending pipetting step is out of sight. Automating library prep is attractive precisely because a machine can hold small volumes and tight ratios steady across a whole plate, but only if the liquid handling is set up to respect how fragile the chemistry is.
This is a practical view of the liquid handling that makes or breaks automated library prep, organized around the steps that fail most often. The kits and chemistries differ; the pipetting risks are shared across almost all of them.
Bead cleanups live or die on ratio accuracy
The paramagnetic bead cleanup, the SPRI step, appears again and again in library prep, and it is where size selection actually happens. The ratio of bead volume to sample volume sets which fragments are kept and which are discarded, so the accuracy of those two volumes is not a nicety, it is the size selection. A bead suspension is also a genuinely difficult liquid: it is viscous, it settles, and the beads themselves must stay evenly distributed to work.
- Handle bead suspensions with a slower flow rate and enough mixing to keep them homogeneous, because a settled suspension delivers the wrong bead mass and ruins the ratio.
- Treat the bead-to-sample ratio as a controlled quantity, tuned and verified, not an approximate pour.
- Give the ethanol washes the same care as any volatile solvent; incomplete or rough washes carry contaminants forward or strip beads off the magnet.
Small volumes and precious enzymes
Library prep is full of small transfers of expensive reagents: ligases, polymerases, adapter and index oligos, often just a few microliters per well. Two things follow. First, the low-volume accuracy problem is everywhere, so the classes for these reagents need to be tuned for small volumes rather than borrowed from larger transfers. Second, waste is costly, so the handling should aim to deliver the intended volume cleanly without over-aspirating expensive mix that then gets thrown away with the tip. Enzyme mixes are often viscous and glycerol-heavy too, which argues for slower flow rates and longer settling than a watery reagent would need.
Normalization: the ratios must actually be equal
Pooling libraries for sequencing assumes each library is represented in proportion to plan, and that assumption rests entirely on the accuracy of the normalization transfers. If some wells get slightly more and others slightly less, the sequencer spends its reads unevenly and low-represented samples can fail to yield enough data. Normalization is therefore another place where consistency across the plate matters more than any single absolute volume: every well should be treated identically, so that the relative amounts come out as intended.
Cross-contamination ends a run quietly
Because a sequencing pool mixes many indexed samples together, cross-contamination between wells is uniquely damaging: a little material carried from one sample into another shows up as misassigned reads that are hard to diagnose after the fact. Fresh tips between samples, gentle dispensing that does not splash, and avoiding techniques that shuttle small excesses between wells are the defenses. In a workflow this sensitive, spending tips to prevent carryover is cheaper than a compromised run.
In library prep, a pipetting error rarely announces itself. It waits, compounds through the cleanups and the pool, and reappears as a failed library or a lost sample when the offending transfer is long forgotten.