Illumina Unveils BioInsight: A New Frontier in Multiomics and AI-Driven Genomic Discovery

In an industry where the sheer volume of biological data threatens to outpace the tools available to interpret it, Illumina has made a definitive move to bridge the divide. Ahead of the 2026 American Society of Human Genetics (ASHG) annual meeting in Montreal, the genomic giant has officially launched the BioInsight platform. This unified software environment is designed to consolidate Illumina’s disparate analysis applications, artificial intelligence models, and curated data resources into a single, cohesive ecosystem, promising to redefine how researchers navigate the complexity of multiomics.

The Genesis of BioInsight: Addressing the Data Bottleneck

The creation of the BioInsight business unit one year ago was not a reactionary measure, but a strategic recognition of a widening "analysis gap." As sequencing technologies have evolved, the output from multiomics—integrating genomics, transcriptomics, proteomics, and epigenomics—has grown exponentially. However, the software required to synthesize these layers of information often remained fragmented, forcing researchers to navigate complex, disconnected workflows that hindered speed and reproducibility.

"With the BioInsight Platform, we are bringing customers one simple place to access the full breadth of our offerings across software, AI, and data," explained Rami Mehio, senior vice president and general manager of the BioInsight business. According to Mehio, the platform is not merely a product launch but a foundational shift in how the company supports the scientific community. By providing a unified interface, Illumina intends to lower the barrier to entry for complex genomic analysis, empowering researchers to spend less time managing data infrastructure and more time deriving biological meaning.

SpliceAI2: The Engine of Precision Genomics

Central to the BioInsight ecosystem is the integration of SpliceAI2, a sophisticated genomic AI model unveiled by Illumina just last week. SpliceAI2 represents a significant leap forward in the ability to identify disease-relevant splice variants—genetic mutations that disrupt the delicate process of RNA splicing, often leading to rare diseases, hereditary cancers, and other physiological disruptions.

The importance of this tool cannot be overstated. In clinical settings, the primary challenge often lies in interpreting "variants of uncertain significance" (VUS). By accurately predicting the functional impact of splice variants, SpliceAI2 provides clinicians and researchers with the evidence needed to classify these variants as pathogenic or benign, a critical step in both diagnosis and therapeutic intervention.

The pedigree of this new model is impressive. Its predecessor, SpliceAI, became a cornerstone of the field, cited in over 3,400 publications and formally incorporated into the guidelines of ClinGen, the NIH-funded resource that defines the clinical relevance of genes and variants. SpliceAI2 builds upon this legacy, offering enhanced precision and a deeper understanding of the molecular mechanics of disease.

A Rigorous Benchmark: Validating the AI Model

The efficacy of SpliceAI2 is detailed in a recent pre-print published by a multidisciplinary team of scientists from Illumina, the University of Oxford, the University of California, San Francisco, and the New York Genome Center. The findings suggest that the model does more than just iterate on its predecessor; it fundamentally changes the approach to transcript reconstruction.

According to the study, SpliceAI2 "reconstructs complete transcripts directly from genomic sequence and considerably improves identification of splice-altering variants and quantification of their effects." This capability was tested against diverse benchmarks, including cryptic splicing, splicing quantitative trait loci (sQTLs), and massively parallel reporter assays.

The data underscores the model’s superiority:

• Disease Identification: In a comparative analysis using matched DNA and phenotype data from Genomics England cohorts, SpliceAI2 identified 17% more disease-relevant splice variants than its competitors.
• Quantification Accuracy: When analyzed against whole-genome and RNA sequencing data from the NIH’s Genotype-Tissue Expression (GTEx) portal, SpliceAI2 demonstrated a 34% improvement in the quantification of splice site usage compared to the next best available model.
• Proteomic Alignment: The researchers noted that SpliceAI2’s predictions align more closely with observed protein effects in the UK Biobank proteomic datasets, suggesting that the model successfully bridges the gap between genomic mutation and functional proteomic outcome.

Ecosystem Integration: DRAGEN 4.6 and Beyond

While SpliceAI2 serves as the flagship AI model, the BioInsight platform is a comprehensive suite. A key component of the launch is the introduction of DRAGEN version 4.6. As the gold standard for secondary analysis, DRAGEN’s latest iteration expands the flexibility of the software, allowing for deployment across a variety of environments, including standard CPU architectures and cloud-based infrastructures. This "software-only" mode is a direct response to customer demand for greater agility, allowing researchers to run high-performance secondary analysis without being tethered to specific hardware configurations.

Furthermore, the platform incorporates:

• BaseSpace and Connected Multiomics: These applications provide the management layer for data storage, sharing, and collaborative analysis.
• BioInsight AI Lab: Acting as the research arm of the unit, the lab focuses on the development of foundational models. By applying multiomic data to train these models, the lab aims to solve complex problems in variant prediction, gene regulation, and beyond.
• Curated Data Resources: The platform offers pre-curated datasets for oncology, rare disease, and population health, allowing researchers to plug their own data into established reference frames immediately.

Flexibility Through a Credits-Based Model

Perhaps the most significant change for existing Illumina customers is the transition to a credits-based access system. Abandoning the rigid, traditional annual subscription model, Illumina is moving toward a usage-based approach. This allows laboratories of all sizes to access the platform’s high-end AI tools and secondary analysis software without the overhead of enterprise-wide licenses.

This model fosters a "try-before-you-invest" culture, enabling scientists to explore new technologies—such as the BioInsight AI Lab’s foundation models—without significant financial risk. It is a strategic move to democratize access to advanced genomic tools, ensuring that small research groups can utilize the same computational power as large-scale pharmaceutical firms.

Implications for the Future of Biological Insight

The launch of BioInsight marks a critical pivot for Illumina. For years, the company was defined primarily by its dominance in the hardware market—the "sequencers on the bench." By transitioning toward a unified software and AI ecosystem, Illumina is signaling that its future lies in the "interpretation of the sequence" rather than just the "reading of the sequence."

The implications for drug discovery and clinical practice are profound. By integrating AI-driven variant prediction (SpliceAI2) with scalable secondary analysis (DRAGEN) and curated multiomic databases, Illumina is creating a pipeline that can compress years of research into months. For the clinician, this means a faster route to diagnosis for patients with rare diseases. For the pharmaceutical researcher, it means a more reliable method for identifying targets and predicting off-target effects during the drug development process.

As the industry converges in Montreal for the 2026 ASHG meeting, the BioInsight platform will likely serve as the centerpiece of the conversation. The transition from raw data to actionable insight has historically been the "dark matter" of genomics—difficult to see, measure, and define. With BioInsight, Illumina is attempting to illuminate this space, providing a standardized language and toolkit that could set the standard for the next decade of genomic research.

Ultimately, the success of the platform will depend on its adoption by the broader community and its ability to continuously integrate new, disparate data types. However, given the initial performance benchmarks of SpliceAI2 and the flexibility of the new credit-based system, Illumina appears well-positioned to lead this next wave of biological discovery. As the company continues to refine its AI capabilities, the promise of true "multiomic" understanding moves from a theoretical goal to a practical, daily reality for researchers worldwide.