Background: Variants in the DMD gene, that encodes the cytoskeletal protein, dystrophin, cause a severe form of dilated cardiomyopathy (DCM) associated with high rates of heart failure, heart transplantation, and ventricular arrhythmias. Improved early detection of individuals at risk is needed. Methods: Genetic testing of 40 male probands with a potential X-linked genetic cause of primary DCM was undertaken using multi-gene panel sequencing, multiplex polymerase chain reaction, and array comparative genomic hybridization. Variant location was assessed with respect to dystrophin isoform patterns and exon usage. Telomere length was evaluated as a marker of myocardial dysfunction in left ventricular tissue and blood. Results: Four pathogenic/likely pathogenic DMD variants were found in 5 probands (5/40: 12.5%). Only one rare variant was identified by gene panel testing with 3 additional multi-exon deletion/duplications found following targeted assays for structural variants. All of the pathogenic/likely pathogenic DMD variants involved dystrophin exons that had percent spliced-in scores >90, indicating high levels of constitutive expression in the human adult heart. Fifteen DMD variant-negative probands (15/40: 37.5%) had variants in autosomal genes including TTN , BAG3 , LMNA , and RBM20 . Myocardial telomere length was reduced in patients with DCM irrespective of genotype. No differences in blood telomere length were observed between genotype-positive family members with/without DCM and controls. Conclusions: Primary genetic testing using multi-gene panels has a low yield and specific assays for structural variants are required if DMD -associated cardiomyopathy is suspected. Distinguishing X-linked causes of DCM from autosomal genes that show sex differences in clinical presentation is crucial for informed family management.
Distinguishing pathogenic variants from non-pathogenic ones remains a major challenge in clinical genetic testing of primary immunodeficiency (PID) patients. Most of the existing mutation pathogenicity prediction tools treat all mutations as homogeneous entities, ignoring the differences in characteristics of different genes, and use the same model for genes in different diseases. In this study, we developed a single nucleotide variant (SNV) pathogenicity prediction tool, Variant Impact Predictor for PIDs (VIPPID; https://mylab.shinyapps.io/VIPPID/), which was tailored for PIDs genes and used a specific model for each of the most prevalent PID known genes. It employed a Conditional Inference Forest model and utilized information of 85 features of SNVs and scores from 20 existing prediction tools. Evaluation of VIPPID showed that it had superior performance (area under the curve = 0.91) over non-specific conventional tools. In addition, we also showed that the gene-specific model outperformed the non-gene-specific models. Our study demonstrated that disease-specific and gene-specific models can improve SNV pathogenicity prediction performance. This observation supports the notion that each feature of mutations in the model can be potentially used, in a new algorithm, to investigate the characteristics and function of the encoded proteins.
Accurate genetic variant interpretation is crucial for disease research and the development of targeted therapies. Artificial intelligence is transforming this field by integrating computational methodologies across structural biology, evolutionary analysis, and multimodal genomic data. This review examines the evolution from traditional rule-based systems and statistical models to contemporary machine learning, deep learning, and protein language models, while addressing critical challenges in variant classification. Key obstacles include data heterogeneity, interpretability, and the persistence of variants of uncertain significance, emphasizing the critical need for explainable artificial intelligence frameworks and more inclusive genomic databases to improve predictive accuracy across diverse populations. Based on the assessment of current variant impact predictors, we propose strategies for enhanced predictor selection, effective multi-omics data integration, and optimized computational workflows. These recommendations aim to enhance variant interpretation accuracy in both research settings and clinical practice, ultimately contributing to advances in personalized medicine.
Gene expression regulation is a sophisticated, multi-stage process, and its robustness is critical to normal cell function and the survival of an organism. Previous studies indicate that differential gene expression at the RNA level is typically attenuated at the protein level through translational regulation. However, how post-transcriptional regulation (PTR) influences expression change during the RNA maturation process remains unclear. In this study, we investigated this by quantifying the magnitude of expression change in precursor RNA and mature RNA across a vast range of different biological conditions. We analyzed bulk tissue RNA sequencing data from 4689 samples, including healthy and diseased tissues from human, chimpanzee, rhesus macaque, and murine sources. We demonstrated that PTR tends to support homeostatic expression of mature RNA by amplifying normal tissue-specific expression of precursor RNA, while reducing expression change of precursor RNA in disease contexts. Our study provides insight into the general influence of PTR on gene expression homeostasis. Our analysis also suggests that intronic reads in RNA-seq studies may contain under-utilized information about disease associations. Additionally, our findings may assist in identifying new disease biomarkers and more effective ways of altering gene expression as a therapeutic strategy.
Accurate predictive modeling of human gene relationships would fundamentally transform our ability to uncover the molecular mechanisms that underpin key biological and disease processes. Recent studies have employed advanced AI techniques to model the complexities of gene networks using large gene expression datasets 1–11 . However, the extent and nature of the biological information these models can learn is not fully understood. Furthermore, the potential for improving model performance by using alternative data types, model architectures, and methodologies remains underexplored. Here, we developed GeneRAIN models by training on a large dataset of 410K human bulk RNA-seq samples, rather than single-cell RNA-seq datasets used by most previous studies. We showed that although the models were trained only on gene expression data, they learned a wide range of biological information well beyond gene expression. We introduced GeneRAIN-vec, a state-of-the-art, multifaceted vectorized representation of genes. Further, we demonstrated the capabilities and broad applicability of this approach by making 4,797 biological attribute predictions for each of 13,030 long non-coding RNAs (62.5 million predictions in total). These achievements stem from various methodological innovations, including experimenting with multiple model architectures and a new ‘Binning-By-Gene’ normalization method. Comprehensive evaluation of our models clearly demonstrated that they significantly outperformed current state-of-the-art models 3,12 . This study improves our understanding of the capabilities of Transformer and self-supervised deep learning when applied to extensive expression data. Our methodological advancements offer crucial insights into refining these techniques. These innovations are set to significantly advance our understanding and exploration of biology.
Mice are widely used as animal models in biomedical research, favored for their small size, ease of breeding, and anatomical and physiological similarities to humans 1,2 . However, discrepancies between mouse gene experimental results and the actual behavior of human genes are not uncommon, despite their shared DNA sequence similarity 3–8 . This suggests that DNA sequence similarity does not always reliably predict functional similarity. On the other hand, RNA-level gene expression could offer additional information about gene function 9,10 . In this study, we undertook characterization and inter-species comparison of human and mouse genes by applying innovative deep learning methodologies to a large dataset of 410K human and 366K mouse bulk RNA-seq samples. This was achieved by using gene representations from our Transformer-based GeneRAIN model 11,12 . These gene representations aggregate information from large gene expression datasets, and provide insights beyond DNA sequence similarity. We identified 2,407 human-mouse homologous genes with high DNA similarity but distinct RNA characteristics, and showed that these genes are more likely to have differing disease/phenotype associations between the two species. Additionally, we found 3,070 homologous genes with low similarity at both the DNA and RNA levels, suggesting the highest risk of discrepancies in study results between the two species. We propose that this approach will support future decision making around whether the mouse will be an appropriate model for studying specific human genes, and whether the results of specific mouse gene studies are likely to be recapitulated in humans. Our methodological innovations offer valuable lessons for future deep learning applications in cross-species omics data. The interspecies gene relationship findings from our study also contribute valuable insights into the gene biology and evolution of the two species.