Tier Classification
How SPACER classifies guide RNAs into quality tiers based on their assay score.
Overview
After computing the assay score (0.0–1.0) for each candidate guide, SPACER assigns it to one of four quality tiers via SpacerTier::from_assay_score(). Tiers provide a quick, actionable classification that helps researchers prioritize guides for experimental validation without needing to interpret raw scores.
Tier Thresholds
| Tier | Assay Score | Label | Color | Meaning |
|---|---|---|---|---|
| Excellent | ≥ 0.80 | Tier 1 | Green | Top-quality guides with strong scores across all evaluated criteria. Recommended as primary candidates for assay development. |
| Good | 0.60–0.79 | Tier 2 | Blue | Solid guides with favorable properties. Suitable for experimental validation, especially when Excellent-tier candidates are limited. |
| Fair | 0.40–0.59 | Tier 3 | Yellow | Acceptable guides with one or more moderate weaknesses. Consider these when higher-tier options are scarce or when targeting a specific region. |
| Poor | < 0.40 | Tier 4 | Red | Low-quality guides with significant concerns. Not recommended for use without compelling biological reasons and additional validation. |
Tier Assignment Logic
Tier assignment is a simple threshold-based classification applied after the assay score is calculated. The canonical implementation is SpacerTier::from_assay_score(), which multiplies the 0.0–1.0 score by 100 and applies integer thresholds:
- The assay score (0.0–1.0) is scaled to 0–100 and compared against tier boundaries
- Boundary values are inclusive: a score of exactly 0.80 is Excellent, 0.60 is Good, and 0.40 is Fair
- Quality flags do not affect tier assignment — a guide with flags can still be Excellent if its assay score is high enough
POLY_T flag is still classified as Excellent-tier, but the flag serves as a warning that researchers should inspect the specific concern before proceeding.Typical Tier Distributions
The proportion of guides in each tier varies depending on the target sequence, enzyme configuration, and enabled scoring components. Typical distributions for well-characterized viral targets:
| Tier | Typical % | Notes |
|---|---|---|
| Excellent | 5–15% | Only the best candidates; more common with Cas13 + ML activity |
| Good | 15–30% | The primary selection pool for most users |
| Fair | 25–40% | Large pool; useful for region-constrained designs |
| Poor | 20–50% | Often guides with extreme GC or homopolymer issues |
Practical Selection Guidance
Tiers are designed to support a practical experimental workflow:
- Start with Excellent: Select 3–5 Excellent-tier guides as primary candidates for your assay
- Supplement with Good: If Excellent-tier guides do not cover your desired target regions, add Good-tier guides to fill gaps
- Use Fair for constrained designs: When you must target a specific conserved region and no higher-tier guides are available, Fair-tier guides are acceptable with additional validation
- Avoid Poor unless necessary: Poor-tier guides should only be used when no alternatives exist and the biological context demands that specific region