Pipeline Options

Configure spacer length, enabled features, and other analysis parameters.

Overview

The SPACER analysis pipeline is configured through the top-level fields in the analyze request. These parameters control the spacer length, which scoring components are active, and what additional analyses (primers, structure prediction) to perform.

Spacer Length

The spacer length determines the number of nucleotides in each candidate guide RNA. Valid ranges depend on the selected enzyme family.

Enzyme	Min Length	Max Length	Default	ML Length
Cas12	18 nt	24 nt	21 nt	21 nt
Cas13	20 nt	28 nt	28 nt	28 nt

Info

When AI activity prediction is enabled, SPACER may constrain the spacer length to match the model's expected input size (21 nt for EasyDesign/Cas12, 28 nt for ADAPT/Cas13).

AI Activity Prediction

Set enable_ml to true to enable ML-predicted activity scoring. SPACER automatically selects the correct model based on the enzyme:

Enzyme	Model	Architecture	Input
Cas12	EasyDesign	Two-stage CNN (classification + regression)	21 nt spacer + 4 nt PAM + flanking context
Cas13	ADAPT	Two-stage CNN (classification + regression)	28 nt guide + flanking context

Structure Prediction

Set enable_structure to true to enable Vienna RNA secondary structure prediction and seed accessibility scoring.

Info

Structure prediction adds computational overhead proportional to sequence length. For sequences over 5 kb, expect additional processing time.

Primer Design

Set primer_mode to "rpa" to generate RPA primer pairs flanking the top-ranked guide target sites. Set to null to disable.

Direct Repeat (crRNA Assembly)

Set direct_repeat to a preset ID (from the capabilities endpoint) to include a crRNA sequence in the results. The direct repeat is prepended to the spacer sequence to form the full crRNA.

Full Configuration Example

json

{
  "sequence": "ATGCTTACGATCGATCG...",
  "enzyme": "cas13",
  "pam": null,
  "spacer_length": 28,
  "enable_ml": true,
  "enable_structure": true,
  "primer_mode": "rpa",
  "direct_repeat": "lwacas13a",
  "max_results": 50,
  "min_score": 40.0
}

Tip

For a fast analysis without optional features, set enable_ml and enable_structure to false and primer_mode to null. This gives the fastest possible analysis using only sequence-based scoring.

Search documentation