variant effect prediction

The paper identifies an upstream DNA motif required for CTCF's enhancer-blocking insulator activity using both new and published data.

The paper identifies vmeQTLs for blood DNA methylation variance in MZ twins, showing gene-environment interactions and cell-type effects relevant to immune disease.

The paper identifies platform-dependent effects of APOL1 variants on plasma levels measured by different proteomic technologies.

This study evaluates sequence-to-omics models to improve reproducibility in genetic fine-mapping across eQTL datasets, showing reduced replication failure rates and enhanced SNP prioritization.

The paper maps crossover events in honeybees using whole genome sequencing to identify genetic determinants of recombination rate variation.

BESTish infers mutation fitness, mutation rate, and environmental strength in clonal hematopoiesis using a diffusion-approximation framework applied to VAF datasets.

This study predicts the impact of all missense variants in ATG9A using structural analysis and variant effect predictors, identifying potentially damaging variants and their molecular mechanisms.

The paper introduces CORAL, a scalable framework to infer mutation rates and spectra across species, revealing variation linked to life-history traits and identifying evolutionary mutational signatures.

EvoPool improves variant effect prediction by integrating evolutionary information into protein language model embeddings using optimal transport.

This paper introduces a pretraining task for genome language models that predicts evolutionary rate, showing improved performance over sequence-only pretraining.

The paper evaluates transcriptomic integration methods for cyanobacterial metabolic models, finding that method selection depends on model configuration and that certain approaches improve flux predictions.

This paper presents a single-cell multiome atlas of the human basal ganglia, linking chromatin states and transcriptomes to cell-type identity and disease-associated variants.

The paper studies mutations and structural variants during double-strand break repair in budding yeast, finding that repair-associated deletions and template switches depend on microhomology and homeology.

The paper creates a single-cell atlas of transcription and chromatin states in the human brain, identifying regulatory programs and linking them to neuropsychiatric disorders.

The paper develops MutaPheno, a model that predicts cancer driver missense mutations using molecular features, achieving strong accuracy and outperforming existing tools.

The paper introduces AlphaTMB, a composite biomarker combining tumor mutation burden with pathogenicity scores, showing improved survival prediction in immunotherapy response.

This paper uses AI to identify and validate non-coding regulatory variants affecting RHD expression, showing strong agreement between predictions and experimental results.

The paper evaluates AlphaGenome's performance in predicting gene expression and chromatin accessibility from personal genomes, finding it excels in accessibility but struggles with gene expression.

This paper develops a deep learning model to predict molecular alterations in never-smoker lung adenocarcinomas from histological images, achieving high accuracy for most features.

BIG-TB evaluates sequence-based machine learning models for predicting antibiotic resistance and variant attribution using *Mycobacterium tuberculosis* genomes, showing DNA-based models outperform protein-based ones.

This study compares genomic prediction methods for multiple time-resolved traits in maize, finding MegaLMM excels in snapshot accuracy while dynamicGP is better for longitudinal accuracy.

PolarBEAR polarizes SNPs and estimates uSFS without outgroup data, showing high accuracy in simulations and human data, with robustness to mutation models.

FusionPath predicts gene fusion pathogenicity using protein structural data and contextual embeddings, outperforming existing methods with AUC scores of 0.95 and 0.87.

PanCNV-Explorer integrates and normalizes copy number variation data across 33 cancer types to provide a unified resource for analyzing CNV frequency and functional impact.

This paper introduces SIGMA, an AI framework that improves CAD risk prediction by integrating whole exome and phenotypic data, achieving higher sensitivity, specificity, and AUC.

Volaria integrates coding and regulatory variation into gene-centered representations for disease outcome prediction using whole-genome sequencing data.

AlphaGenome predicts functional genomic measurements from DNA sequences with high resolution and accuracy across multiple modalities, outperforming existing methods in variant effect prediction.

The paper evaluates genetic-constraint metrics across the human noncoding genome and identifies regions where they underperform due to model bias and low signal-to-noise ratios.

The paper optimizes sensitivity and specificity in targeted and whole-exome sequencing using reference-standard DNA mixtures, identifying optimal pipelines and VAF thresholds to reduce false positives and improve detection.

Lilace improves variant effect estimation in FACS-based DMS by providing uncertainty quantification and better FDR control.