PromoterAtlas: decoding regulatory sequences across Gammaproteobacteria using a transformer model

2026-05-15 · Source: Machine learning : nature.com subject feeds · Field: Science & Research — Life Sciences & Biology, Artificial Intelligence & Machine Learning, Synthetic Biology · Depth: Expert, quick

Summary

PromoterAtlas is a 1.8 million parameter transformer model designed to decode regulatory sequences across Gammaproteobacteria. Trained on 9 million regulatory sequences from 3371 species, this model overcomes limitations of previous bacterial promoter prediction tools, which were often constrained by small datasets and species-specific training. PromoterAtlas accurately recognizes diverse regulatory elements, including ribosomal binding sites, various bacterial promoters, transcription factor binding sites, and terminators across different species. The model also functions as a whole-genome promoter annotation tool for Gammaproteobacteria, with validations supporting predictions for different sigma (σ) factors. Its embeddings reflect cross-species evolutionary relationships, clustering promoters by σ factor identity, and effectively predict transcription and translation levels.

Key takeaway

For synthetic biologists and bacterial geneticists working with Gammaproteobacteria, PromoterAtlas offers a robust tool for understanding and engineering bacterial regulatory sequences. You should consider integrating this model for comprehensive whole-genome promoter annotation and for predicting gene expression levels. Its ability to decode diverse regulatory elements across species can significantly enhance your experimental design and lead to more precise genetic modifications.

Key insights

PromoterAtlas is a transformer model that decodes bacterial regulatory sequences across thousands of Gammaproteobacteria species.

Principles

• Large, diverse datasets improve bacterial sequence analysis.
• Transformer embeddings can reveal evolutionary relationships.
• Regulatory sequence information predicts expression levels.

Method

PromoterAtlas, a 1.8M parameter transformer, was trained on 9M regulatory sequences from 3371 gammaproteobacterial species to recognize diverse regulatory elements and annotate whole genomes.

In practice

• Annotate whole genomes for Gammaproteobacteria.
• Predict transcription and translation levels.
• Engineer bacterial regulatory sequences.

Topics

• PromoterAtlas
• Transformer Model
• Gammaproteobacteria
• Regulatory Sequence Analysis
• Bacterial Promoter Prediction

Best for: AI Scientist, Research Scientist

Related on AIssential

• Decoding sequence determinants of gene expression in diverse cellular and disease states — Decima predicts cell-type-specific gene expression from DNA sequence, enabling precise regulatory mechanism and variant effect analysis.
• Leap in DNA Synthesis Slashes Time to Build New Genetic Sequences — Sidewinder enables rapid, high-fidelity DNA synthesis, resolving a key bottleneck for generative AI in biology.
• SpecGP as a transformer-based model for predicting energy-adaptable structural spectra of glycopeptides — SpecGP uses a transformer architecture to predict glycopeptide spectra and retention times, enhancing isomeric discrimination and identification.
• [R] Genomic Large Language Models — Genomic LLMs can identify functional similarities in DNA beyond sequence alignment, indicating learned regulatory patterns.
• Diffusion model generating regulatory DNAs — DNA-Diffusion is a diffusion model generating functional regulatory DNA sequences conditional on cell type.
• OCOO-T : A Simple and Scalable Virtual Cell Model for Transcriptional Perturbation Response Prediction — OCOO-T uses a simple Transformer and flow matching to directly denoise gene expression for perturbation prediction.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.