Helixer
Helixer is an ab initio gene prediction tool for identifying protein-coding genes directly from raw eukaryotic DNA sequences. It combines deep learning-based base-wise predictions with a Hidden Markov Model to generate structural gene models without requiring additional experimental evidence such as RNA-seq data or protein alignments. Helixer is available as an online service and as open-source software for local installation.
Key benefits
Ab initio prediction of protein-coding genes from genomic DNA
Combines deep learning with Hidden Markov Model-based postprocessing
Does not require RNA-seq data or protein homology evidence
Suitable for structural annotation of eukaryotic genomes
Available via web interface and GitHub
Applications
Structural gene annotation of newly sequenced eukaryotic genomes
Prediction of primary protein-coding gene models
Genome annotation workflows without additional experimental evidence
Comparative genomics and functional genomics projects
Preparation of gene models for downstream biological interpretation
Intended use
Helixer is intended for genome annotation researchers, bioinformaticians, eukaryotic genomics researchers, and plant scientists who need to predict protein-coding genes from genomic DNA sequences. It is particularly suited for projects where RNA-seq or protein evidence is unavailable, incomplete, or not intended to be used.
Contact:
Website