The rule

A vrRNA guide segment is a tandem array of 5-nt units. Each unit is Y + NN + ZZ: NN is a dinucleotide copied out of your target, and Y and ZZ are fixed published backbone.

• The target is read pair 2, skip 1. Dinucleotide i is target[3i .. 3i+1]; the base after it is never contacted.
• So a target is 3 × units − 1 nt long, and the last unit has no skip base after it.
• The dinucleotide goes in verbatim — no complement, no reversal. Recognition is geometric, not canonical Watson–Crick pairing.
• Case is a reading convention only: UPPERCASE for the NN you program, lowercase for the fixed scaffold. It carries no sequence information.

Scaffold tables

Rendered from the same constants the design engine uses (app/core/systems.js), so what you read here is literally what the tool applies. Highlighted cells are the published backbone variants.

Worked examples

The four examples from the design deck. Each is re-derived by this tool from its target and compared against the published guide, every time this page loads. A mismatch here means the tool is broken — do not use it.

What the published work establishes

Colour key

Do

• Swap your target’s dinucleotides into the NN positions.
• Keep every backbone base exactly as published.
• Pick either strand — there is no PAM and no strand bias.
• Append the system-specific 3′ tail at the cloning stage.
• Read the finished vrRNA back against the scaffold before ordering.

Do not

• Touch the GGY backbone motifs. Point mutations there abolished silencing.
• Add or remove bases — the 1-nt skip spacing must be preserved exactly.
• Complement or reverse the dinucleotides.
• Mix tails between systems. They are not interchangeable.