variant effect prediction

DecoyOpt improves EC number prediction by distinguishing active enzymes from decoys, achieving high accuracy and low false-positive rates.

This study shows that excluding low-frequency variants and indels in fine-mapping leads to false discoveries and introduces SuSiE-OUT to address this issue.

This paper introduces a variant filtering framework called SMART filtering to improve diagnostic outcomes in genome sequencing by prioritizing non-coding and regulatory variants.

The paper develops machine learning models to predict artemisinin resistance in Plasmodium falciparum K13 mutants using physicochemical and structural features of mutations.

This study evaluates genomic prediction methods for growth traits in black soldier fly larvae across different diets and generations, finding that dominance effects and SNP density significantly impact predictive ability.

MELD predicts SNV functional effects by directly training on wild-type versus mutant differences, outperforming post-hoc in-silico mutagenesis methods with fewer parameters.

This paper uses automated patch-clamp assays to reclassify SCN1A missense variants, changing classifications for 50% of tested variants.

SARDINE uses EWAS summary statistics to estimate epigenetic clock-phenotype associations without individual-level data, revealing heterogeneity in biomarker associations with aging and diseases.

This paper introduces MirrorDNA, a DNA language model with a strand-equivariant tokenizer that improves prediction consistency across DNA strands.

GIBD detects IBD segments in pangenome graphs and annotates them with structural variants, outperforming existing methods in accuracy and speed.

memVar links human membrane protein variants to molecular contexts and disease, revealing enrichment of pathogenic variants in transmembrane regions and interaction sites.

PALM, a protein language model, predicts domain, functional site, and variant effects with high accuracy using curated sequence data and outperforms existing models in multiple benchmarks.

The paper assesses the transportability of mouse gene knockout models to human complex phenotypes using GPAT framework, finding evidence for 12 phenotypes but limited for others.

The paper evaluates how small reference sample sizes affect substructure estimation and proposes bias corrections for improved accuracy.

CellReg3D integrates DNA language models and 3D genome structure to model cell-type-specific gene regulation, achieving state-of-the-art performance in predicting epigenetic marks and gene expression.

MicrobeRT is a framework of fine-tuned language models for metagenomic analysis, achieving strong performance in taxonomic classification, antibiotic resistance detection, and pathogenicity assessment.

This study identifies genetic polymorphisms in NKG2C, CD16a, and HLA-E that influence NKG2C expression and ADCC capacity in HCMV-infected individuals.

SeverYX is an ensemble framework that classifies genetic variants into severity tiers, achieving high concordance with ClinVar and improving predictive accuracy through threshold optimization and additional predictors.

LRCQ estimates phenome-wide heritability enrichment of genomic loci using GWAS summary statistics, identifying enriched regions and their functional annotations.

This paper introduces Perturbscape, a framework that uses perturbation-induced gene programs to identify disease-relevant genes and pathways, showing higher heritability informativeness than steady-state approaches.

The paper shows that neural RNA splicing models rely on reading frame information that the splicing machinery cannot access, leading to errors in predicting poison exons.

This study functionally characterizes Kv7.5 channel variants, identifying gain- and loss-of-function effects and proposing a z-score classification for pathogenicity.

The paper identifies rare founder variants in a population and develops a platform for variant prioritization, showing that some pathogenic variants may be missed by conventional methods.

The paper presents a curated dataset of 1,413 LNP cytotoxicity measurements to examine structure-activity relationships and develop computational models of lipid nanoparticle cytotoxicity.

FastaGuard is a FASTA preflight QC tool that detects structural issues in assemblies and stops pipelines before expensive analysis, with high accuracy in identifying faults.

CNVision is a computational framework that automates the interpretation of intragenic copy number variants by standardizing input parsing, transcript mapping, and reading-frame classification.

This paper adapts ACMG/AMP variant classification criteria for the non-protein-coding gene RMRP, resulting in a framework for classifying variants in cartilage-hair hypoplasia.

This paper improves variant-effect prediction in sequence-to-function models using synthetic MPRA data generated from distilled model predictions.

The paper introduces eQIF, an unsupervised machine-learning method that integrates Enformer predictions to prioritize gene-regulatory variants, showing strong performance in predicting eQTLs and GWAS variants.

This paper introduces PGL, a Bayesian regression method for microbiome feature selection that uses phylogenetic trees to improve detection of exposure-associated ASVs.