SeqTrainer helper

Get an SBOL annotation prompt

Fill in the benchmark paths and scan settings below. Seq Design Buddy will prepare a command you can copy into PowerShell on the machine where SeqTrainer is installed.

This form creates text only. It does not upload your GenBank file, run SeqTrainer, or validate the resulting SBOL file. Predictions are computational annotations.

Benchmark details

Fields marked * are needed to produce a useful command.

Input GenBank file *

Example: $env:USERPROFILE\\Downloads\\pAN1717_cyan.gb

The plasmid .gb file that SeqTrainer will scan.

Model bundle folder *

Example: outputs\\models\\dnabert2_kaggle_best

The trained benchmark bundle containing compatible model files.

Validation threshold *

Example: 0.80

Use the threshold selected on validation data, usually from the benchmark manifest.

Annotated GenBank output *

Example: outputs\\annotations\\pAN1717\\annotated.gb

Where the annotated .gb file will be written.

Optional benchmark files

Leave these blank when the model bundle already contains the needed files.

Separate checkpoint path

Example: outputs\\models\\dnabert2_kaggle_best\\checkpoint.pt

Use this when the trained checkpoint is outside the model bundle.

Benchmark manifest path

Example: outputs\\models\\dnabert2_kaggle_best\\manifest.json

Records model settings and the validation-selected threshold.

Scan and output options

Window size

Example: 300

Optional scan window length. The manifest or SeqTrainer default is used when blank.

Step size

Example: 25

Distance between scan windows, in bases.

Merge distance

Example: 25

Optional distance for merging nearby positive windows.

Predictions CSV

Example: outputs\\annotations\\pAN1717\\predictions.csv

Optional per-window scores and coordinates output.

Annotation manifest

Example: outputs\\annotations\\pAN1717\\manifest.json

Optional provenance and settings JSON output.

SBOL3 output

Example: outputs\\annotations\\pAN1717\\annotated.nt

Optional SBOL3 N-Triples output for later validation.

SBOL2 RDF/XML output

Example: outputs\\annotations\\pAN1717\\annotated.rdf

Optional RDF/XML compatibility output for SBOL Canvas import.

Scan both strandsAdd --scan-both-strands to inspect forward and reverse-complement sequences.Preserve existing featuresKeep known GenBank features in the annotated output.Clean previous outputRemove this run's previous outputs before writing new files.Open output folderAsk SeqTrainer to open the output folder after completion.

Generated prompt

Review the command, then use the copy button to move it into PowerShell.

PowerShell command

seqtrainer annotate promoters <input-genbank-path> --model-family dnabert2 --model-bundle <model-bundle-path> --threshold <validation-threshold> --output <annotated-genbank-output> --scan-both-strands --preserve-existing-features --clean-output --open-output-folder

Expected files

SeqTrainer should write the annotated GenBank file and any optional CSV, manifest, or SBOL outputs you selected. Validate SBOL output before opening it in Canvas.