Joseph A. Marsh

The paper shows that variant effect predictors influence ClinVar classifications, creating circularity in variant classification and prediction.

This study identifies ATRX deficiency as a driver of aberrant type I interferon signaling via cGAS-dependent transcriptional dysregulation in patients with cerebral white matter disease.

The paper uses deep mutational scanning to distinguish disease mechanisms in ERCC2 variants, finding that TTD variants have lower fitness than XP variants.

The paper compares variant effect predictors and MAVEs to understand their agreement and disagreement in assessing genetic variant impacts.

The paper identifies ER-liquid condensates that sequester FAM134B/C and RhoA, affecting cell morphology and cancer invasion.

The paper introduces MES, a metric to evaluate variant effect predictors and MAVEs based on clinical evidence yield, showing discrepancies with AUROC.

The paper identifies proteins where variant effect predictors disagree, suggesting these are prioritized for experimental validation to improve variant interpretation.

This study determines the structural morphology and formation of the human exocyst complex, revealing an extended architecture with minimal intersubunit interaction and a single nucleation site.

The paper quantifies signalling effects of CTNNB1 mutations across cancer types and finds weakly activating mutations correlate with worse prognosis in hepatocellular carcinoma.

This study shows that SS18::SSX can drive transcription independently of BAF complex interaction and relies on EP300 for gene activation, suggesting new therapeutic targets for synovial sarcoma.

This study combines MAVEs and population-free VEPs to improve variant classification across ancestries in hereditary cancer genes, reclassifying over 90% of variants of uncertain significance.

The paper identifies molecular domains in the U8 snoRNA affected by LCC mutations and shows how U8 function impacts stability through a post-transcriptional mechanism.

The paper develops a protein structure-based mLOF score to predict loss-of-function mechanisms and estimates the prevalence of dominant-negative and gain-of-function mechanisms in Mendelian disease genes.

The paper introduces acmgscaler, an R package and Colab notebook for calibrating gene-level variant effect scores within the ACMG/AMP framework using MAVEs and VEPs.

The paper evaluates variant effect predictors and disease mechanisms in intrinsically disordered proteins, finding reduced sensitivity for pathogenic variants in these regions and highlighting the need for disorder-informed prediction strategies.

The paper develops a method to estimate enzymatic activities from genetic variants using fitness assays and combines allele effects to predict pathogenicity in recessive diseases.

The paper shows that disease-associated variants in alternative protein isoforms can cause structural mutations, emphasizing the need to consider tissue-specific expression and full-length transcripts in variant interpretation.

This paper introduces SPDV, a spatial distance metric that outperforms half of existing variant effect predictors in identifying dominant disease variants, particularly for gain-of-function missense variants.

The paper investigates epitope effects in pQTL studies, finding that only 12% or less of cis-pQTL results are influenced by missense-mediated epitope effects.

This study evaluates 97 variant effect predictors using DMS experiments and finds strong correlation between their performance in functional assays and clinical variant classification.

This paper evaluates non-coding variants' impact on TF binding using AlphaFold 3 and FoldX, showing structural modeling can provide interpretable insights despite lower accuracy than sequence-based methods.

This study evaluates 70 variant effect predictors for RyR1 missense variants and introduces a structural metric, SPDV, to assign ACMG/AMP evidence levels for clinical interpretation.

This study evaluates 35 variant effect predictors across 963 human genes, finding gene-level performance variability and identifying intrinsic disorder as a key factor affecting AUROC metrics.

The paper introduces a statistical model to predict dominant-negative, gain-of-function, and loss-of-function mechanisms of protein-altering mutations in human genes and validates its performance on unseen data.

This study determines the crystal structure of human PADI6, revealing structural differences that suggest it lacks catalytic activity compared to other PADIs.

This study shows ALDH1A3 and ACSS2 coordinate acetaldehyde-acetyl-CoA metabolism with histone acetylation in melanoma, contributing to transcriptional heterogeneity and drug resistance.

The paper uses deep mutational scanning to quantify PAX6:DNA binding and predict clinical outcomes of variants, finding most pathogenic mutations affect DNA binding.

This study evaluates ancestry bias in variant effect predictors and finds disparities in their performance across different populations.

The paper identifies TUBB4B mutations causing distinct ciliopathic diseases by affecting microtubule dynamics and cilia formation in humans and mice.

The paper analyzes cancer mutations in protein structures to identify novel driver genes by distinguishing mutational patterns in tumour suppressors and oncogenes.