Illumina Launches SpliceAI2 Genomics AI Model, Lifting Disease-Variant Detection by 17%

What happened

SpliceAI2 was developed by Illumina's BioInsight AI Lab, reports say, with one core job: interpreting splice variants — predicting how a genetic variant changes RNA splicing, and thus whether it is disease-related. On rare-disease benchmark datasets, SpliceAI2 detected 17% more disease-related variants than comparable models.

Compared with the original 2019 SpliceAI, the new model was trained on more than 100 times as much data. Illumina has released the code and weights on GitHub and Hugging Face for academic, non-commercial use.

Key facts

Released by: Illumina (BioInsight AI Lab), October 8, as reported.

Capability: predicts how genetic variants alter RNA splicing to flag pathogenic variants.

Benchmark: 17% more disease-related variants detected than peer models on rare-disease datasets.

Data scale: trained on 100x more data than the original 2019 SpliceAI.

Open: code and weights on GitHub/Hugging Face under an academic, non-commercial license.