Glossary
The terms this site uses most, in alphabetical order, defined once. Every definition is drawn from prose already on the site, and each entry says where the term is explained in context.
- adaptive sampling
- Enrichment performed by the sequencer rather than at the bench. The instrument begins reading a molecule, decides whether it belongs to the target region, and ejects it from the pore if it does not. Real-time enrichment
- basecalling
- Converting the raw electrical signal a pore produces into a sequence of bases. It is done by a machine-learning model, which is why the quality scores it reports are estimates from that model rather than measurements. Measuring sequencing accuracy
- cleavage-ligation cycling CLC
- A one-pot library preparation that combines Type-IIS cleavage with adapter ligation in a single tube, in place of the separate shearing, end-repair and ligation steps, and supports barcoded multiplexing. Bench chemistry
- CYP2D6
- A pharmacogene that determines how patients metabolize a large fraction of prescribed drugs. It is structurally complex, which is why resolving it needs long reads. Clinical application
- CYP2D7
- The near-identical pseudogene lying beside CYP2D6. The two are similar enough that short reads cannot reliably be assigned to one rather than the other, which is the difficulty the long-read work exists to address. Clinical application
- diplotype
- The pair of haplotypes a person carries at a locus, one per chromosome copy. For a pharmacogene this is the clinically actionable answer, and it is only as good as the confidence attached to the reads supporting it. Clinical application
- end reason
- The annotation a nanopore writes for every read recording why that read stopped. On a molecule of known length it marks whether the read finished, and a quality filter cannot stand in for it. Most pipelines discard it when converting to FASTQ or BAM. End reason filtering
- flow cell
- The consumable holding the array of pores a nanopore run sequences through. Its condition changes over a run, which is one reason run-level quality is not constant. Measuring sequencing accuracy
- Golden Gate Type-IIS Golden Gate assembly
- A cloning method that uses Type-IIS restriction enzymes, which cut outside their own recognition site, so several DNA fragments can be joined in a defined order in one reaction without leaving a scar. Reference materials
- haplotype
- The set of variants carried together on one copy of a chromosome. Because a long read comes from a single molecule, the variants it spans are known to sit on the same copy. Clinical application
- long-read sequencing
- Sequencing that reads a single DNA molecule end to end, for thousands of bases at a time, rather than assembling a picture from many short fragments. The lab works on Oxford Nanopore and PacBio HiFi. Research overview
- pharmacogene
- A gene whose variants change how a patient responds to a drug, in dose, efficacy or adverse reaction. Clinical application
- plasmid standard
- A circular DNA molecule of known, independently verified sequence, built in the lab and sequenced alongside real samples so that any disagreement between the read and the template is attributable to the instrument. Measurement and standards
- read-end integrity
- Whether the bases at the ends of a read are really there. A read can fall short of its true end, which matters when the whole point is to score a read against a template of known length. Error characterization
- SMA-seq Single-Molecule Accuracy sequencing
- Measuring a sequencer's error rate per read against templates whose sequence is already known, rather than against a reference assembled from reads. The standards are built in-house so that ground truth is a physical molecule. Measurement and standards
Terms were selected by counting how many pages actually use each one, not by taste: everything above appears on at least three pages of this site, which is where a reader arriving cold meets it repeatedly.