Search

The authors have applied an eight-tissue rat multi-omic dataset to analyze the gene regulatory landscape of endurance exercise training, identifying tissue-specific regulatory patterns involved in muscle function, metabolism and immune response.

In a randomized study involving 9 general cardiologists and 107 real-world patient cases, assistance from a specifically tailored large language model resulted in preferable responses on complex case management compared to physicians alone, as rated by specialist cardiologists using a multidimensional scoring rubric.

A transformer-based vision system for cardiac MRI offers a generalizable and data-efficient system for diverse tasks in cardiology, offering clear advantages for contextualizing human cardiovascular disease and diagnosing disease.

Evo 2 is an artificial intelligence-based biological foundation model trained on 9 trillion DNA base pairs spanning all domains of life that predicts functional properties from genomic sequences and provides a rich generative model for researchers in biology.

Wang, Tang and colleagues develop the low-signal signed iterative random forest pipeline to investigate epistasis in the genetic control of cardiac hypertrophy, identifying epistatic variants near CCDC141, IGF1R, TTN and TNKS loci, and show that hypertrophy in induced pluripotent stem cell-derived cardiomyocytes is nonadditively influenced by interactions among CCDC141, TTN and IGF1R.

Via an integrative modelling approach that combines population and clinical trial data, the authors find that polygenic risk score-based screening would reduce premature mortality across seven commonly screened conditions.

There is a genetic component to the risk of severe COVID-19, but the genetic effects are difficult to separate from social constructs that covary with genetic ancestry. To address this, the authors identify determinants of COVID-19 severity using admixture mapping, viral phylodynamics, and host immune and metagenomic sequencing.

A global network of researchers was formed to investigate the role of human genetics in SARS-CoV-2 infection and COVID-19 severity; this paper reports 13 genome-wide significant loci and potentially actionable mechanisms in response to infection.

The availability of labelled training data is one of the practical obstacles towards wide application of machine learning models in medicine. Here the authors develop a weakly supervised deep learning model for the classification of aortic malformations using unlabelled cardiac MRI sequences from the UK biobank.

A video-based deep learning algorithm—EchoNet-Dynamic—accurately identifies subtle changes in ejection fraction and classifies heart failure with reduced ejection fraction using information from multiple cardiac cycles.

Using well-calibrated statistical methods the authors jointly analyze Undiagnosed Diseases Network genomes, identifying known and novel disease genes. Software is publicly available to support future cross-cohort rare disease discovery efforts.

The genetic and pathogenetic basis of heart failure is incompletely understood. Here, the authors present a high-fidelity tissue collection from rapidly preserved failing and non-failing control hearts which are used for eQTL mapping and network analysis, resulting in the prioritization of PPP1R3A as a heart failure gene.

Regular exercise promotes overall health and prevents non-communicable diseases, but the adaptation mechanisms are unclear. Here, the authors perform a meta-analysis to reveal time-specific patterns of the acute and long-term exercise response in human skeletal muscle, and identify sex- and age-specific changes.

A streamlined sequencing process enables identification of disease-causing variants in the clinic within 8 hours.

DeepFlow uses deep learning to extract cardiac flow phenotypes from phase-contrast magnetic resonance images. Genome-wide analyses of cardiac flow traits in UK Biobank highlight the contribution of connective tissue genes to aortic valve function.

An analysis of 24,202 critical cases of COVID-19 identifies potentially druggable targets in inflammatory signalling (JAK1), monocyte–macrophage activation and endothelial permeability (PDE4A), immunometabolism (SLC2A5 and AK5), and host factors required for viral entry and replication (TMPRSS2 and RAB2A).

Precision medicine is a strategy for tailoring clinical decision making to the underlying genetic causes of disease. This Review describes how, despite the straightforward overall principles of precision medicine, adopting it responsibly into clinical practice will require many technical and conceptual hurdles to be overcome. Such challenges include optimized sequencing strategies, clinically focused bioinformatics pipelines and reliable metrics for the disease causality of genetic variants.

DNA variant calling methods based on deep neural networks can use local haplotyping information with long-reads to improve genotyping accuracy, however this increases computational complexity. Here the authors develop an approximate haplotagging method that simplifies the process and enables state-of-the-art variant calling performance with multiple sequencing platforms.

Mendelian randomization is a popular method to detect causal relationships between traits, but can be confounded by instances of horizontal pleiotropy. Here, the authors present a Mendelian randomization workflow which includes causal discovery analysis and filtering of genetic instruments based on their conditional independencies.

APJ is shown to be a bifunctional receptor for both mechanical stretch and the endogenous peptide apelin, a finding that is important for the development of APJ agonists to treat heart failure.

Over 90% of human whole-genome sequencing has been performed using instruments from two companies, Illumina and Complete Genomics. Lam et al. sequence the same DNA samples with both instruments and compare their performance for calling insertions, deletions and single-nucleotide variants.

A set of three papers in Nature reports a new proteomics resource from the UK Biobank and initial analysis of common and rare genetic variant associations with plasma protein levels.

MILTON uses phenotype information in the UK Biobank to identify clinical biomarkers and other quantitative traits that characterize diseases. It then constructs augmented cohorts by predicting undiagnosed individuals, improving power to discover gene–disease relationships.

Temporal multi-omic analysis of tissues from rats undergoing up to eight weeks of endurance exercise training reveals widespread shared, tissue-specific and sex-specific changes, including immune, metabolic, stress response and mitochondrial pathways.

In this Review, Weldy and Ashley summarize the biology of heart failure (HF), the mechanisms of action of HF drugs, and both the Mendelian genetics of cardiomyopathies and the complex genetics of HF. Finally, they present a roadmap for the future of precision medicine in HF.

Single-nucleus ATAC-seq characterization of chromatin accessibility in human coronary artery disease samples identifies cell-type- and state-specific regulatory mechanisms underlying disease risk, highlighting the roles of TBX2 and PRDM16.

A diagnostic tool based on blood RNA-seq is shown to identify causal genes and variants linked to clinical phenotypes in individuals with rare diseases for which whole-exome genetic sequencing was uninformative.

It is known that exercise influences many human traits, but not which tissues and genes are most important. This study connects transcriptome data collected across 15 tissues during exercise training in rats as part of the Molecular Transducers of Physical Activity Consortium with human data to identify traits with similar tissue specific gene expression signatures to exercise.

Cardiac hypertrophy is associated with a decrease in expression of the adult isoform of the molecular motor myosin heavy chain (α-MHC) and the induction of expression of its fetal isoform (β-MHC). Here the authors reveal the mechanism regulating this switch in expression, which impairs heart function. Cardiac stress in adult hearts reactivates the developmental chromatin-modifying complex Brg1/BAF, which interacts with histone deacetylase and poly (ADP ribose) polymerase to induce a pathological α-MHC-to-β-MHC shift.

Whole-genome sequencing, transcriptome-wide association and fine-mapping analyses in over 7,000 individuals with critical COVID-19 are used to identify 16 independent variants that are associated with severe illness in COVID-19.

Using a multi-omics approach, the authors examine the molecular drivers of sexual dimorphism in the subcutaneous adipose tissue from sedentary and endurance-trained rats. These data provide a valuable resource for adipose tissue-related research.

Protein-truncating variants (PTVs) are predicted to significantly affect a gene’s function and, thus, human traits. Here, DeBoever et al. systematically analyze PTVs in more than 300,000 individuals across 135 phenotypes and identify 27 associations between PTVs and medical conditions.

Many large research initiatives have cumulatively enrolled thousands of patients with a range of complex medical issues but no clear genetic etiology. However, it is unclear how researchers, institutions and funders should manage the data and relationships with those participants who remain undiagnosed when these studies end. In this Comment, we outline the current literature relevant to post-study obligations in clinical genomics research and discuss the application of current guidelines to research with undiagnosed participants.

To build a more efficient, equitable and sustainable approach to rare disease research in the United States, we must prioritize integrated research infrastructure and approaches that focus on understanding connections across rare diseases.

In this Review, Roden and co-workers describe how multiplexed assays of variant effects can be used for high-throughput functional assessment of nearly all coding variants in a target sequence to improve variant annotation for cardiovascular-related genes and to resolve the problem of classification as variants of uncertain significance. They also discuss how variant effect predictors can be integrated with multiplexed methods to inform cardiovascular genomic medicine.

Relatives of patients with amyotrophic lateral sclerosis have an unexpectedly high incidence of schizophrenia. Here, the authors show a genetic link between the two conditions, suggesting shared neurobiological mechanisms.

The GREGoR consortium provides foundational resources and substrates for the future of rare disease genomics.

Das et al. present a novel Bayesian approach called expression Quantitative Trait enhancer Loci (eQTeL), which effectively integrates genetic and epigenetic information to identify combination of regulatory genomic variants underlying expression variance. Using various functional data, the authors show the variants identified by eQTeL are likely to be causal.

Here, a long noncoding RNA, termed Mhrt, is identified in the loci of myosin heavy chain (Myh) genes in mice and shown to be capable of suppressing cardiomyopathy in the animals, as well as being repressed in diseased human hearts.