EasyDesign Model
Two-stage CNN for predicting Cas12 guide RNA on-target activity.
Overview
The EasyDesign model is a two-stage convolutional neural network trained to predict the on-target activity of Cas12 guide RNAs. SPACER automatically selects this model when AI activity prediction is enabled for Cas12 analyses.
Huang Y, Zhang W, Cheng Y, Wang Y, Wei C, Sun Y, Yang L, He S (2024). “Deep learning enhancing guide RNA design for CRISPR/Cas12a-based diagnostics” iMeta.
Info
EasyDesign is the ML model for Cas12. For Cas13 guides, SPACER uses the ADAPT model instead.
Model Architecture
EasyDesign uses a two-stage CNN cascade:
- Stage 1 (Classification): Predicts whether a guide is active or inactive
- Stage 2 (Regression): For guides classified as active, predicts a continuous activity score
Input Encoding
| Parameter | Value |
|---|---|
| Spacer length | 21 nt |
| PAM | 4 nt (prepended internally) |
| Flanking context | 10 nt upstream + 10 nt downstream |
| Encoding | One-hot with 10 channels per position (5 target + 5 guide, with gap support) |
Scoring Integration
When AI activity prediction is enabled for a Cas12 analysis, SPACER runs EasyDesign on each candidate guide. The predicted activity score is incorporated into the composite score via the ML adjustment component.
Warning
When EasyDesign is active, the spacer length is constrained to 21 nt to match the model's training input. This is set automatically when you enable AI activity prediction for Cas12.
Performance
| Metric | Value |
|---|---|
| Applicable enzyme | Cas12 |
| Required spacer length | 21 nt |
| Output range | Continuous activity score |
| Inference | Batched via ONNX Runtime |