AI
Genomic models, agents, and AI-assisted tooling for genomics and variant interpretation.
AI in Genomics & Biomedical Research
The AI tab tracks the intersection of artificial intelligence with human genomics and biomedical discovery. Key trajectories:
-
Genomic foundation models (DNA/RNA): From BPE-based tokenisers (DNABERT, Ji et al. PMID 33542176, Bioinformatics 2021) to reverse-complement equivariant architectures (HyenaDNA, Nguyen et al. NeurIPS 2023; Caduceus, Schiff et al. NeurIPS 2024) to whole-genome autoregressive models (Evo 2, Brixi et al. PMID 39664581, Science 2025; Genos, Arora et al. bioRxiv 2025). Scaling trend: kilo-base → mega-base context windows, single-species → cross-species generalization.
-
Pathology foundation models: 2024 explosion — self-supervised whole-slide image models (UNI, Chen et al. Nat Med 2024; Virchow, Vorontsov et al. Nat Med 2024; Prov-GigaPath, Xu et al. PMID 38931993, Nature 2024) → vision-language alignment for multimodal interpretation (CONCH, Lu et al. Nat Med 2024; TITAN, Ding et al. Nat Med 2025) → knowledge-enhanced architectures (KEEP, Li et al. PMID 39972922, Nat Med 2026).
-
AI for GWAS: Image-based phenotyping via CNN/U-Net (Haas et al. PMID 34957434, Cell Genom 2021; Khurshid et al. PMID 36944631, Nat Commun 2023) → self-supervised contrastive learning from raw images (iGWAS, Kirchler et al. Nat Genet 2024) → Transformer-based association methods (InsightGWAS, Song et al. bioRxiv 2025) → AI-enhanced post-GWAS interpretation (PoPS, Weeks et al. PMID 37106029, Nat Genet 2023).
-
Autonomous AI agents: From assistive diagnostic tools (AI-MARRVEL, Mao et al. PMID 39631886, NEJM AI 2024) to fully autonomous scientific discovery — hypothesis generation, experimental design, and execution (AI Scientist, Lu et al. ICML 2025; Robin, Qiu et al. Nat Biotechnol 2026; Co-Scientist, Gottweis et al. Nature 2026).
-
Regulatory genomics models: From short-range sequence-to-function prediction (DeepSEA, Zhou & Troyanskaya PMID 26301841, Nat Methods 2015) to cell-type-specific, long-range expression models integrating epigenomic context (Enformer, Avsec et al. PMID 34316034, Nat Methods 2021; Borzoi, Linder et al. PMID 39704929, Nat Genet 2025).
Contents - AI
- Agent : Review - 2 , Tools - 11
- Auto_research : Tools - 9
- Benchmark : Review - 3
- Clinical_EHR : Tools - 1
- Coding : CLI - 2 , Extension - 1 , IDE - 4
- Datasets :
- Clinical_EHR : Dataset - 2
- Health_Survey : Dataset - 2
- Physiological_Signals : Dataset - 1
- Drug_discovery : Tools - 1
- GWAS :
- Imaging_GWAS : Method - 5 , Tools - 2
- Methods : Tools - 4
- Phenotyping : Tools - 3
- Post_GWAS : Tools - 1
- Review : Perspective - 1 , Review - 2
- Genomic_language_model : Model - 12
- Imaging : Tools - 10
- Major_model_series : Model series - 12
- Multimodal : Tools - 1
- Protein : Tools - 1
- Regulatory_model : Model - 6
- Single_cell : Tools - 1