How to use guntingAI

Target lookup

Search

1. Select a reference or private model.
2. Choose the Cas system and enter the expected PAM pattern.
3. Paste one strict 20-nucleotide spacer per line.
4. Set mismatch, result-count, and model-distance limits where needed.
5. Run the search and review the result summary, coordinate map, PAM, strand, annotation, mismatches, and model distance.

Candidate generation

Design

1. Select a model, Cas system, and PAM.
2. Choose a target type: gene name, gene ID, genomic interval, or pasted DNA.
3. Provide the target and optionally restrict the preferred region.
4. Set GC and off-target screening preferences.
5. Review ranked candidate spacers, their observed PAMs, target location, risk signals, and coordinate map.

Panel compatibility

Multiplex

1. Enter 2–30 guide spacers, one per line.
2. Add optional labels separated from each spacer by a tab, comma, or vertical bar.
3. Select the model and PAM rules.
4. Run the panel analysis.
5. Review individual guide scores, pair compatibility, recommended combinations, shared target signals, and mapped best targets.

Account-restricted reference

Private Models

1. Open Workspace → Private Models.
2. Provide a model name, organism or source, and optional provenance note.
3. Upload or paste a FASTA file within your account genome limit.
4. Optionally add GFF3, GTF, BED, or guntingAI TSV annotation.
5. Ensure annotation sequence IDs match the first token of the corresponding FASTA headers.
6. Queue training and track the job under Recent Jobs.

Workspace and records

Account usage

• Recent Jobs: track Search, Design, Multiplex, and private training jobs.
• Private Models: view, publish-request, or delete owned models and submit new training.
• Account Provenance: review the institution, affiliation, position, and identity associated with model ownership.
• Account Limits: view queue and training allowances.
• Saved results: open records, view target maps, download JSON, or remove records.