Helixer: ab initio prediction of primary eukaryotic gene models combining deep learning and a hidden Markov model

2025/11/24 by Felix Holst, Anthony Bolger, Felicitas Kindel +11 · 1 voice · 7 citations
Biochemistry, Genetics and Molecular Biology · #Bioinformatics and Genomic Networks #Genomics and Phylogenetic Studies #Machine Learning in Bioinformatics

paper · pdf · doi:10.1038/s41592-025-02939-1

openalex publication_date 2025/11/24 · openalex created_date 2025/11/25 · openalex updated_date 2026/07/31

Abstract

The accurate identification of genes is vital for understanding biological function, yet this remains challenging across many newly sequenced or less-studied species. Here we present Helixer, an artificial intelligence-based tool for ab initio gene prediction that delivers highly accurate gene models across fungal, plant, vertebrate and invertebrate genomes. Unlike traditional methods, Helixer operates without requiring additional experimental data such as RNA sequencing, making it broadly applicable to diverse species. We show that Helixer's pretrained models achieve accuracy on par with or exceeding current tools, producing gene annotations that closely match expert-curated references across multiple evaluation metrics. Its design enables immediate use on genomes without retraining, providing an efficient, accessible solution for genome annotation in both research and applied settings. The tool is available as an open-source software for local installation via GitHub. An online web interface is also available as well as through the Galaxy ToolShed.

Citations

• Highly accurate protein structure prediction with AlphaFold
• Effective gene expression prediction from sequence by integrating long-range interactions
• Tiberius: end-to-end deep learning with an HMM for gene prediction
• A full genome assembly reveals drought stress effects on gene expression and metabolite profiles in blackcurrant (Ribes nigrum L.)
• Language Models are Few-Shot Learners
• GFF Utilities: GffRead and GffCompare
• Deep Learning for Symbolic Mathematics
• Pseudo-Labeling and Confirmation Bias in Deep Semi-Supervised Learning
• Attention Is All You Need
• Semi-supervised Sequence Learning
• Phytozome: a comparative platform for green plant genomics
• Gene identification in novel eukaryotic genomes by self-training algorithm

Cited by

• A draft genome assembly of the agricultural pest Leucoptera coffeella and analysis of its dsRNA processing machinery is a key step toward RNAi-based biopesticides in Lepidoptera
• Water mass specific genes dominate the Southern Ocean microbiome
• A conserved structural logic underlies sensor–helper NLR communication in the NRC immune receptor network
• Chromosome-Scale Genome Assemblies of Duckweeds Provide Insights Into Genomic Plasticity, Aquatic Adaptation, and Morphological Reduction
• Pangenome analysis reveals the evolutionary dynamics of repeat-based holocentromeres
• Transposable elements in the era of data science EMBO course: a call for open, accessible, and collaborative TE resources
• LitCTL1: A novel C-type lectin involved in the mucosal and cellular immunity of the common periwinkle Littorina littorea

Discussions

• This could transform genome reseach f.e. to be used in plant #breeding and for faster identification of genes that influence drought or disease resistance. Described in @natmethods.nature.com doi.o [bsky, 1 points, 0 comments]

Related

• Accurate ab initio gene prediction in eukaryotes with Tiberius in multiple clades
• Tiberius: end-to-end deep learning with an HMM for gene prediction
• Tiberius: End-to-End Deep Learning with an HMM for Gene Prediction
• GeneCAD: Plant Genome Annotation with a DNA Foundation Model
• White paper: The Helix Pathogenicity Prediction Platform