variant effect prediction

EpiReasoner integrates AI for phenotype-to-genotype reasoning in plant epidermal development, identifying a major QTL for stomatal density in tomatoes.

The paper identifies genetic drivers of clinically relevant traits in Candida parapsilosis using genome-wide association studies and machine learning models.

HAUDI improves polygenic score construction for admixed populations by efficiently extending the GAUDI method to multi-way admixture with faster computation and better performance.

The paper combines enriched case data with population-based case-control analyses using a likelihood-ratio methodology to classify variants in breast cancer susceptibility genes.

This paper evaluates Evo 2's ability to predict pathogenicity and mechanisms of cardiomyopathy variants, showing high accuracy and structural feature identification.

HiFiMAP provides high-resolution IBD mapping at SNP level, outperforming window-based methods and detecting novel associations in UK Biobank and brain imaging data.

This paper benchmarks synthetic data generation methods for clinical transcriptomic cancer data, finding no single method outperforms all metrics, with Gaussian Copula showing the most balanced performance.

This study argues that fixing the intercept in LDSC improves SNP heritability estimation by reducing bias and variance, but LDSC remains biased under population stratification.

SarcDBase is a tool that extracts and interprets genetic alterations from tumor sequencing data, showing high agreement with curated diagnoses in most cases.

The paper uses AlphaGenome to assess the transcriptional impact of somatic mutations in aging, finding their effects are much smaller than endogenous aging transcriptional changes.

mafsmith reimplements vcf2maf in Rust, achieving faster and more efficient conversion of VCF files to MAF with identical output across multiple datasets.

This paper improves variant effect prediction by modifying protein language models with sparse mechanistic features, achieving significant performance gains across multiple datasets.

Bio-BLIP is a multimodal architecture that integrates DNA, genes, proteins, and text to improve variant interpretation without task-specific fine-tuning, achieving better performance in variant prioritization and target gene prediction.

mehari predicts variant effects with a strict HGVS-first approach, achieving high speed and accuracy compared to existing tools.

SAFE-h2 is a tool that filters SNPs to improve heritability estimation by removing those with negative contributions, revealing phenotypic variance from both additive and non-additive allelic effects.

This paper introduces DeepUMQA-Global, a deep learning framework for estimating protein structure model accuracy, which outperforms existing methods in correlation metrics and consensus rankings.

The paper identifies polygenic risk factors and rare variants in unexplained familial cancers using whole genome sequencing and family history data.

The paper introduces BDXR, a method for antimicrobial resistance prediction, achieving high accuracy across six species-drug pairs in the CAMDA 2026 challenge.

This paper develops a germline polygenic score (PRSagg) to predict prostate cancer aggressiveness and validates it across multiple datasets.

The paper compares two photosynthesis models using light response data and finds the Johnson and Berry formulation reduces global gross primary productivity estimates by 8%.

The paper introduces VIS-seq, an image-based method to assess variant effects on cellular phenotypes, revealing multidimensional variant impacts across different diseases.

The paper analyzes genome sequencing data from UK Biobank to assess the penetrance of COL1A1/2 pathogenic variants in OI, finding lower penetrance than expected.

This paper trains a neural network to correct FoldX's free energy predictions for protein folding and binding, significantly improving accuracy for single and double mutations and epistasis.

The paper introduces a model-free permutation method to estimate P-values in genetic studies, reducing computational demands while maintaining accuracy.

The paper characterizes a CACNA1A variant in a patient with mixed clinical symptoms, showing conflicting in vitro and in vivo effects that highlight the importance of cellular context.

This study evaluates Shapeit2 duoHMM's reliability and accuracy in inferring recombination events on whole genome sequence data compared to SNP arrays.

The paper introduces IRES-TrAPPr, a method to quantify CDI and IRES activity, revealing that viral IRESes from warm-blooded hosts are more structurally stable and many DNA-based IRES candidates have negligible activity.

This paper presents a protocol for minigene assays to validate splice-altering variants using locus-specific design, gene synthesis, and long-read sequencing.

STARMAP maps protein structural features to regulatory and cellular phenotypes using natural genetic variation across cancer cell lines.

The paper introduces MR-SimSS, a method to correct biases in two-sample MR studies by simulating independent data subsets from GWAS summary statistics, improving causal inference accuracy.