Illumina launches SpliceAI2 for interpreting splice variants: AI shrinks the uninterpretable genome a little further

## Illumina launches SpliceAI2 On October 8, Illumina released SpliceAI2, a genomic AI model that predicts how genetic variants alter RNA splicing. This is a key step in interpreting "variants of uncertain significance" in rare disease, hereditary cancer testing and drug discovery. **Key points** - **Performance**: On rare disease research datasets, SpliceAI2 identified 17% more disease-relevant variants than other splicing models, helping researchers catch splice variants that would otherwise be missed. - **Data scale**: Its training dataset is more than 100 times that of the previous-generation SpliceAI (released in 2019). The original model has been cited in more than 3,400 papers and was incorporated into splice variant interpretation recommendations from ClinGen, the clinical genomics standards body. - **A suite**: SpliceAI2, PromoterAI and PrimateAI-3D form a genomic AI suite covering three types of variant effect: splicing, promoter and missense. Together, they can double the number of variants detected as "predicted to have biological impact". - **Direction**: Illumina's BioInsight AI lab says it is using AI to "systematically shrink the part of the genome that remains uninterpretable", as genomics moves from "being able to measure" to "being able to understand". > Source: [read the original](https://www.unite.ai/illumina-launches-spliceai2-model-for-splice-variant-interpretation/). This text is a summary for reference only.