Gene Prediction

Upload a genome file to predict protein-coding genes. All fields have sensible defaults.

Input File (provide either a file or a pasted sequence)
Accepted formats: FASTA (.fa, .fasta, .fna, .ffn), GenBank (.gb, .gbk), EMBL (.embl). Gzip compressed files are also supported.
— or paste sequence below —
Paste a FASTA-formatted nucleotide sequence.
Output Settings
GFF3 is recommended for most use cases. BED for genome browser tracks. GenBank format for NCBI submission.
Genetic code used for translating DNA to protein. Default: Auto (automatic detection).
Enable to receive additional outputs alongside the main prediction result. ORF-flower map visualization is enabled by default.
Prediction Parameters
SVM score cutoff for gene prediction. Higher values increase precision, lower values increase recall. Range: 0–1.
Minimum length of a putative gene in base pairs. Shorter values may detect small ORFs but increase false positives.
Prediction strategy. Single (default) for isolated genomes; Meta for metagenomic assemblies.
Topology assumed for sequences without topology annotation (plain FASTA). Default: Linear. Choose Circular for complete genomes/plasmids.
Training Files Upload the model parameter cache file produced by Clovers self-training. Using it when predicting closely related species can greatly reduce model training time. (optional)
Custom SVM model file. Leave empty to use the built-in model.
Custom RBS training file. Leave empty to use the built-in RBS model.
Equivalent Command 
clovers -i input.fa -o result.gff -f gff -l 90 -g auto -s 0.499 -p single
This command mirrors the current form selections for local execution.