Full archive
Bioinformatics & AI
Week of 6 May 2026
11 articles
Geno'X Veille — genox-veille.fr
Key takeaways this week
- ►AgentIA (Nature 2026) — agentic system for rare diseases — traceable step-by-step reasoning: resolution of complex cases with explainable audit trail; a milestone in clinical utility of diagnostic AI.
- ►AlphaGenome (Nature 2026) — unified prediction of regulatory variant effects on chromatin, splicing and expression: 10× more precise than previous models across 50+ functional modalities.
- ►PheMART (Nature Biomed Eng 2026) — phenotype-specific missense variant prediction via contrastive learning + EHR: outperforms AlphaMissense and CADD across 4,179 phenotypes; 5.1M variant database available.
- ►HiFi long-read RNA-seq EJHG 2026 — +8 diagnoses in 25 patients vs short-read RNA-seq: variant phasing on RNA, complete isoforms, intron retention — emerging gold standard for splicing variants.
- ►Franklin leads ACMG/AMP benchmark on 151 real clinical VCFs (Bioinformatics 2026): reassess InterVar if still used as sole classification tool.
- ►LLMs in genetics 2026 — Gemini 2.5 Pro resolves 6 rare disease cases without fine-tuning; Med-PaLM 2 discovers new genetic factor for deafness validated by knock-in mouse.
- ►New PVS1_RNA + BP7_S criteria (Clin Chem 2026) — ClinGen SVI Splicing Subgroup: SpliceAI prediction + RNA functional assay combination essential to reclassify BRCA1/BRCA2 splicing VUS.
Category:
Filters··Min. score
11 articles of 11
Sort
Rare diseases — diagnosis by agentic AI system
PubMed★ Top pick
⭐ À la une
An agentic system for rare disease diagnosis with traceable reasoning
Rare diseases — diagnosis by agentic AI system
0
LLM appliedClinical pipeline
Nature, 2026· MarRead
Regulatory variants — functional effect prediction on chromatin, splicing and expression
PubMed★ Top pick
Advancing regulatory variant effect prediction with AlphaGenome
Regulatory variants — functional effect prediction on chromatin, splicing and expression
0
Pathogenicity predictionNew tool
Nature, 2026· JanRead
Missense variants — phenotype-specific pathogenicity prediction
PubMed★ Top pick
Phenotypic prediction of missense variants via deep contrastive learning
Missense variants — phenotype-specific pathogenicity prediction
0
Pathogenicity predictionNew tool
Nature Biomedical Engineering, 2026· AprRead
Long-read
PubMed★ Top pick
HiFi long-read RNA sequencing enhances clinical diagnostics in rare disorders
Rare diseases with suspected splicing variants — LRS vs SRS differential diagnostic yield
0
Long-readLong-read sequencingDiagnostic RNA-seq
European Journal of Human Genetics, 2026· MarRead
Long-read
PubMedPopulation-level structural variant characterization using pangenome graphs
Structural variants (SV) — population-level catalog via pangenome
0
Long-readSV caller
Nature Genetics, 2026· MarRead
Rare pediatric genetic diseases — undiagnosed after WES/short-read WGS
PubMedDiagnostic utility of clinical genome reanalysis in rare pediatric disorders using long-read sequencing
Rare pediatric genetic diseases — undiagnosed after WES/short-read WGS
0
Long-read sequencingGenomic reanalysis
HGG Advances, 2026· AprRead
Clinical pipeline
PubMedScaling genomic reanalysis to unlock diagnoses and transform rare disease care
Rare genetic diseases — optimization of diagnostic yield through systematic reanalysis
0
Clinical pipelineGenomic reanalysisDiagnostic yield
HGG Advances, 2026· AprRead
Long-read
PubMedTargeted long-read RNA sequencing for rare disease diagnosis and variant interpretation
Rare genetic diseases — variant interpretation by targeted long-read RNA-seq
0
Long-readLong-read sequencingDiagnostic RNA-seq
Science Advances, 2026· AprRead
Algorithm benchmark
PubMedComprehensive evaluation of ACMG/AMP-based variant classification tools
Mendelian diseases — ACMG/AMP classification tool benchmark in real clinical context
0
Algorithm benchmarkACMG classificationBenchmark
Bioinformatics (Oxford), 2026· FebRead
Algorithm benchmark
PubMedSplicing Predictions, Splicing Assays, and Variant Classification Using ACMG/AMP Guidelines: Challenges Observed with BRCA1 and BRCA2 Variants
HBOC — reclassification of BRCA1/BRCA2 splicing variants
0
Algorithm benchmarkACMG classificationVUS reclassified
Clinical Chemistry, 2026· FebRead
Rare diseases — LLM use for genetic diagnosis and new gene discovery
PubMedGenetic Diagnosis and Discovery Enabled by Large Language Models
Rare diseases — LLM use for genetic diagnosis and new gene discovery
0
LLM applied
2026· AprRead
References and sources
- Rare diseases — diagnosis by agentic AI system. Nature, 2026. PMID 41708847. Score 9/10. https://pubmed.ncbi.nlm.nih.gov/41708847/
- Regulatory variants — functional effect prediction on chromatin, splicing and expression. Nature, 2026. PMID 41606153. Score 9/10. https://pubmed.ncbi.nlm.nih.gov/41606153/
- Missense variants — phenotype-specific pathogenicity prediction. Nature Biomedical Engineering, 2026. PMID 41981312. Score 9/10. https://pubmed.ncbi.nlm.nih.gov/41981312/
- Rare diseases with suspected splicing variants — LRS vs SRS differential diagnostic yield. European Journal of Human Genetics, 2026. PMID 41807732. Score 9/10. https://pubmed.ncbi.nlm.nih.gov/41807732/
- Structural variants (SV) — population-level catalog via pangenome. Nature Genetics, 2026. PMID 41807798. Score 8/10. https://pubmed.ncbi.nlm.nih.gov/41807798/
- Rare pediatric genetic diseases — undiagnosed after WES/short-read WGS. HGG Advances, 2026. PMID 42050932. Score 8/10. https://pubmed.ncbi.nlm.nih.gov/42050932/
- Rare genetic diseases — optimization of diagnostic yield through systematic reanalysis. HGG Advances, 2026. PMID 41715921. Score 8/10. https://pubmed.ncbi.nlm.nih.gov/41715921/
- Mendelian diseases — ACMG/AMP classification tool benchmark in real clinical context. Bioinformatics (Oxford), 2026. PMID 41688871. Score 7/10. https://pubmed.ncbi.nlm.nih.gov/41688871/
- HBOC — reclassification of BRCA1/BRCA2 splicing variants. Clinical Chemistry, 2026. PMID 41499258. Score 7/10. https://pubmed.ncbi.nlm.nih.gov/41499258/
- Rare diseases — LLM use for genetic diagnosis and new gene discovery. 2026. PMID 41655254. Score 7/10. https://pubmed.ncbi.nlm.nih.gov/41655254/
- Rare genetic diseases — variant interpretation by targeted long-read RNA-seq. Science Advances, 2026. PMID 41984969. Score 8/10. https://pubmed.ncbi.nlm.nih.gov/41984969/
