Illumina SpliceAI2: genomic AI model for splice variant interpretation
Illumina's BioInsight AI Lab released the successor to the 2019 SpliceAI, trained on 100x more data, predicting how genetic variants alter RNA splicing and finding 17% more disease-relevant variants in a rare-disease dataset.
Why it matters
Continues Illumina's stack of genomic AI (with PromoterAI and PrimateAI-3D); the original SpliceAI has 3,400+ citations and ClinGen adoption, so the upgrade inherits real-world clinical trust.
What you could build with it
A rare-disease diagnostics startup could integrate SpliceAI2 into its variant prioritization pipeline to re-analyze previously inconclusive exome cases, converting years of 'variants of uncertain significance' into billable diagnostic answers for families.
Does it hold up?
Strong inherited trust but too early for SpliceAI2 itself — the predecessor has 3,400+ citations and ClinGen adoption; SpliceAI2 claims 17% more disease-relevant variants on a rare-disease dataset with no independent replication yet.
Built with Illumina SpliceAI2
- Illumina leadership perspectives: genomic AIillumina.com · Background on Illumina's BioInsight AI Lab genomic AI program: the original SpliceAI's approach to cryptic splice mutations and its reported 95% splice-junction accuracy at release.
- Illumina unveils PromoterAI, a groundbreaking algorithm for rare disease diagnosisinvestor.illumina.com · Illumina's PromoterAI announcement details how the BioInsight AI Lab's models combine — PromoterAI, PrimateAI-3D and SpliceAI doubling diagnostic yield — the same stack SpliceAI2 now upgrades.
Learn more
First spotted on article: source.
More AI models
OpenAI publishes 722 AI-generated math manuscripts from an unreleased modelOpenAI released 722 mathematical manuscripts in 372 result families to GitHub under Apache-2.0, produced by an…models · JEV 0.74Mistral Large 4 (Le Chonk): 1.05T-param open-weight MoEMistral's new flagship: a 1.05T-parameter mixture-of-experts multimodal model with 49B active params and 1M-token…models · JEV 0.72ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71Mistral Forge: enterprise platform for training and continuously improving proprietary modelsMistral AI launched Forge (Sep 28) — a full-lifecycle model training platform (pre-training, SFT, DPO/ODPO, RL…models · JEV 0.69DeepSeek V4-Flash official API: public beta with upgraded agent capabilities, Responses API and Codex supportDeepSeek's official V4-Flash is now a public beta API with massively upgraded agent capabilities - benchmark scores…models · JEV 0.69
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.