Skip to the content

The life of a test · 01 Design

Design the assay before the first plate.

Primers designed and checked against the genome and the variants of the breed, the positive control built and confirmed, and both pinned into the test version.

Primer design

Primers that miss the variants.

Known variants are masked before a primer is placed, so a carrier never reads as a minor allele.

Primer design around a target, with known variants marked under the candidate primers
Primer pairs ranked around the target, each variant under a primer shown with its frequency.
  • Designed around the target with the lab's own rules
  • Variants by breed masked, with their frequencies
  • Saved to the primer library with orders and lots

Primer check

Every pair checked against the genome.

Where a pair binds, what it would amplify and which variants sit under it, checked again by itself when a new variant release arrives.

A primer pair checked against the genome, with its products and the variants under it
One product as intended, and the variant that sits under the reverse primer.
  • Every product the pair could make, likely or possible
  • Dimers and hairpins at the lab's conditions
  • A new variant release flags the tests it touches

Cloning

Controls built, then proven.

Restriction, Gateway, homology and Golden Gate cloning, each junction checked as the reaction is built, and the construct confirmed by its own reads.

A Golden Gate reaction with each junction checked
A Golden Gate assembly, its overhangs and the fidelity of every junction.
  • Golden Gate junction fidelity from published ligation data
  • Construct history that can be assembled again
  • A confirmed construct pinned as the test's positive control

Construct confirmation

Every disagreement listed.

Reads assembled against the plan, every difference named, and thin coverage blocking confirmation until it is sequenced again.

Sanger reads assembled against a construct plan, with disagreements listed
A contig against its plan, with one disagreement to resolve.
  • Each conflict resolved with a reason
  • Coverage that is too thin is said, not passed
  • The confirmed version is the one the test pins