Snpxpress: ML prediction of promoter SNP impact on gene expression
SNPxPress is a computational tool that predicts how promoter SNPs affect gene expression to support regulatory variant prioritization in crop improvement. It integrates promoter sequences, SNP catalogs, motif disruption scores, chromatin accessibility profiles, and RNA-seq data to produce ranked SNP lists and interpretable expression-change predictions. The modular software is designed to fit into existing bioinformatics workflows and reduce experimental burden by enabling in silico prioritization.
Overview
SNPxPress is a computational tool designed to predict the impact of promoter single nucleotide polymorphisms (SNPs) on gene expression. It supports regulatory variant prioritization by providing a decision-making framework focused on promoter SNPs and their expected effects on expression.
By integrating promoter sequences, SNP catalogs, motif disruption scores, chromatin accessibility profiles, and RNA-seq data, SNPxPress delivers interpretable predictions of expression changes. The software is modular and intended to enhance existing bioinformatics workflows, enabling efficient prioritization of regulatory variants.
Technical specifications
SNPxPress uses gradient-boosted trees for feature selection and interpretable deep learning to model higher-order interactions. This combination produces ranked SNP lists, disrupted motif maps, and predicted expression shifts.
The model is initially trained on Arabidopsis data due to comprehensive data availability, and is then applied to soybean seed composition traits, including oil/protein balance and fatty acid metabolism. The approach is fully in silico, aiming to minimize experimental costs while providing transparent and interpretable predictions.
Technology readiness level
Currently at Technology Readiness Level (TRL) 3, SNPxPress is a validated computational prototype. The project is planned in four phases over 32 weeks, culminating in a reporting kit with validation guidelines. Each phase includes deliverables intended to ensure alignment and transparency, providing a pathway for future experimental validation and extension to other crops.
About ADC Consulting
ADC, formerly known as Amsterdam Data Collective, is a European AI transformation consultancy that helps organizations move from AI ambition to practical, scalable implementations. The firm provides end-to-end support, from defining AI strategy to building and deploying data and AI platforms, including advanced decision science models, LLMs, and agent-based systems. Their approach integrates strategy, delivery, and design—often in collaboration with their design studio, CLEVER°FRANKE—to ensure that systems perform at scale and deliver measurable value. By combining scientific rigor with deep industry knowledge, they bridge the gap between stakeholder requirements and technical execution.
ADC serves a global client base across sectors such as financial services, life sciences, retail, transportation, and public society. Their work enables organizations to solve complex challenges, such as accelerating drug development, optimizing pricing strategies, and conducting AI feasibility studies. By embedding governance, change management, and human-centered design, the firm helps clients overcome the limitations of pilot projects and integrate AI into real-world operations. Their team of over 100 professionals, operating from offices in the Netherlands, Denmark, and Sweden, works alongside global leaders to drive responsible, data-driven transformation.
delivered to your inbox.