Volume 58 Issue 5, May 2026

Genetic prediction of type 1 diabetes is one of the most successful for complex traits. A machine learning approach now improves this further and discovers multiple non-linear locus–locus interactions and molecular subclusters with differing clinical features.

Most cases of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) are genetically sporadic. By detecting low-frequency somatic mutations across postmortem brains, we show that rare somatic mutations are spatially restricted to focal regions yet drive widespread degeneration. This supports a ‘somatic-spread’ model, according to which local genetic lesions initiate widespread neurodegeneration.

By mapping chromatin accessibility over transposable elements (TEs) across normal hematopoietic cells and the various cell types of primary acute myeloid leukemia (AML), we identified TE subfamilies with altered chromatin state in leukemia stem cells. Signatures of TE chromatin accessibility were able to predict clinical outcomes in cohorts of patients with AML, linking repetitive DNA elements to stemness.

This study highlights the Wheat Spatial Omics Consortium, which aims to build a spatiotemporal single-cell atlas of wheat in response to different treatments, thereby contributing to the development of sustainable and climate-resilient wheat.

The transcriptional interference model has been instrumental in shaping our understanding of gene regulation. However, given that its effects are largely absent from genome-wide data, this Perspective reexamines the underlying mechanisms and suggests how they may be resolved at most loci.

Mendelian randomization is widely used but relies on specific assumptions that are rarely systematically assessed. This Perspective argues that researchers should rigorously test these assumptions through careful empirical analysis that incorporates a broader range of data.

This Review discusses the principles and applications of Strand-seq and highlights its ability to assign genetic variants based on their maternal or paternal origin and potential to contribute to chromosome-length phasing.

This review examines how ancient DNA has transformed our understanding of human adaptation by enabling genetic changes to be tracked directly through time, revealing how past shifts in environment, diet and disease shaped present-day human variation.

Targeted sequencing of neurodegenerative disease genes and transcriptome-wide sequencing in postmortem brain and spinal cord samples from individuals with amyotrophic lateral sclerosis or frontotemporal dementia identify somatic mutations as likely drivers of disease pathology.

Multi-ancestry and ancestry-stratified genome-wide analyses identify genetic variants associated with refractive error and enable construction of an enhanced polygenic predictor incorporating functional annotations.

Genome-wide analyses identify genetic loci and plasma proteins associated with polycystic ovary syndrome (PCOS). This study highlights the hormonal and metabolic foundations of the disease and explores the impact of polygenic risk for PCOS in both sexes.

Multi-ancestry genome-wide association analyses coupled with multi-omics integration identify new risk loci for endometriosis and adenomyosis, shed light on underlying molecular mechanisms and suggest potential therapeutic interventions.

Genome-wide association and fine-mapping analyses of type 1 diabetes (T1D) identify multiple genetic risk signals. Furthermore, a machine learning model, T1GRS, improves the prediction of T1D in individuals with complex risk profiles and identifies genetic subgroups.

This study explores the clonal architecture of aplastic anemia across age using single-cell approaches. Somatic inactivation of specific human leukocyte antigen risk alleles is a frequent event and often occurs in multiple independent events.

This study identifies distinct transposable element subfamilies as genetic determinants of stemness properties in normal and leukemic stem populations with clinical implications for patients with acute myeloid leukemia.

Profiling of pleural mesothelioma samples from 91 patients identifies four DNA methylation subtypes that correlate with response to immune checkpoint inhibition and survival

This paper identifies the ‘master regulators’ that maintain cell transcriptional states in diffuse midline gliomas and shows that targeting them in combination could represent a promising therapeutic avenue for the disease.

SpaMosaic is a mosaic integration method for spatial omics data that enables cross-modality and cross-batch integration using contrastive learning and graph neural networks.

INSPIRE addresses challenges to integrating diverse spatial transcriptomics datasets by combining deep learning with non-negative matrix factorization, revealing shared and context-specific spatial gene programs and tissue organization across scales.

Telomere-to-telomere genome assemblies for two diploid and four tetraploid peanut varieties provide insights into peanut genome organization, evolutionary dynamics and phenotypic differentiation.

A pangenome of the Rosa subgenus constructed from genomes of 26 Rosa accessions, including 23 newly assembled, highlights widespread structural variations and key genes associated with ornamental traits for rose breeding.

A population-scale super-pangenome constructed from 138 reference-grade genome assemblies, representing seven extant Citrullus species, highlights genomic variation associated with fruit-quality traits and accelerates watermelon breeding.