Clovers — Protein-Coding Gene Recognition System v1.1.0
Clovers (Clustering ORFs Via Ensemble Refinement Strategy) is a high-performance ab initio gene finder that revisits and reformulates the geometric model-based paradigm pioneered by the ZCURVE family with a novel technical route. Clovers delivers state-of-the-art performance, enables the de novo discovery of overlapping genes that are difficult for conventional tools to detect, and runs faster with a reasonable memory footprint.
Clovers accepts the following file formats. Compressed variants (gzip) are also supported.
.fa .fasta .fna .ffn.gb .gbk .gbff.emblgbk) — Standard GenBank flat file format with annotations. gbk-full) — Extended GenBank format with complete feature details, including original sequence and all attributes. Compatible with NCBI submission pipelines.| Field | Description |
|---|---|
Upload File |
Upload a genome file in FASTA (.fa, .fasta, .fna, .ffn), GenBank (.gb, .gbk), or EMBL (.embl) format. Gzip compressed files are also supported. |
Paste FASTA Sequences |
Paste a FASTA-formatted nucleotide sequence directly into the textarea. Mutually exclusive with file upload — providing one will disable the other. |
Example |
Loads a built-in example genome for quick testing, pre-populated from static/fa/example.fa. |
| Parameter | Default | Description |
|---|---|---|
Output format |
GFF3 | Format of the prediction result. Options: GFF3 (General Feature Format, recommended), BED (Browser Extensible Data), GenBank (gbk), and GenBank (full) (gbk-full, with complete feature details for NCBI submission). |
Translation table |
auto | Genetic code used for translating DNA to protein. Options: Auto — automatic detection (default), 1 (Standard), 4 (Mycoplasma), 11 (Bacterial, Archaeal), 15 (Mitochondrial/Protozoan), 16 (Mitochondrial/Vertebrate), 25 (Candidate Division SR1). |
| Parameter | Default | Description |
|---|---|---|
Score threshold |
0.5 | SVM score cutoff for gene prediction. Higher values increase precision (fewer false positives), lower values increase recall (more predictions). Range: 0–1. |
Min gene length (bp) |
90 bp | Minimum length of a putative gene in base pairs. Shorter values may detect small ORFs but increase false positives. Minimum allowed: 30 bp. |
Procedure |
single | Prediction strategy (-p). single (default) for isolated single genomes; meta for metagenomic assemblies. |
Default topology |
Linear | Topology assumed for sequences without topology annotation (plain FASTA). linear (default) treats every sequence as linear; choose circular (-c) for complete genomes and plasmids. |
| Field | Description |
|---|---|
SVM training file |
Upload a custom SVM model file to override the built-in model. Leave empty to use the default. |
RBS training file |
Upload a custom RBS training file to override the built-in RBS model. Leave empty to use the default. |
| Option | Description |
|---|---|
Protein translations (.faa) |
When enabled, a .faa file containing the translated protein sequences will be generated alongside the main prediction result. |
Nucleotide sequences (.fna) |
When enabled, a .fna file containing the nucleotide sequences of predicted genes will be generated alongside the main prediction result. |
ORF-flower map visualization |
When enabled (default), an ORF-flower map visualization of the predicted genes is generated from the Clovers debug output and shown on the result page. |
Clovers can be used directly from the command line for batch processing and pipeline integration:
clovers -i input.fasta -o output.gff -f gff -c -g 11 -l 90 -s 0.499
For a complete list of options:
clovers --help
Additional command-line flags include BLAST+ integration (-b, -e, -C), frameshift correction (-F), and GOL-Reporter overlap analysis (-O, -A, -D).
If Clovers contributes to your research, please cite:
Zhang Z, You Z, Lin Y*, Gao F*. High-performance ab initio gene prediction with Clovers.
For questions, bug reports, or collaboration inquiries: