Cancer Genomics and Diagnostics

This paper introduces MMRL, a framework for learning representations of somatic missense mutations across cancer types, showing improved prognostic discrimination and identification of mutation subtypes.

This study evaluates the impact of an individual-specific pangenome on somatic variant calling in a pediatric AML patient, finding reduced false positives and altered phylogenetic branching.

ctDNA-BERT scores individual bisulfite reads for cell-free DNA methylation analysis, achieving high accuracy in cancer staging without cohort-specific retraining.

This study shows that chromosomal instability (CIN) drives rapid genome changes in cancer cells, generating new copy number alterations each cell cycle and increasing genomic complexity.

The paper creates a pan-cancer catalog of mutation patterns at transcription factor binding sites across 15 tissue types, revealing enrichment in nearly one third of TF-binding site pairs.

Clonucopya links tumor clonal architecture to drug prioritization by integrating subclonal inference and drug ranking tools in an end-to-end workflow.

This study compares whole-genome sequences of adolescent and young adult solid cancers to older adult cancers, finding shared genetic drivers but age-related differences in mutational processes and germline variants.

This paper introduces biopsy-guided methods for phylogenetic reconstruction from longitudinal single-cell copy number profiles, showing improved accuracy over pooled methods in simulated data and demonstrating feasibility in patient-derived data.

This study identifies RER deficiency as a new DNA repair defect in cancer, detectable through mutational signatures and linked to therapeutic vulnerabilities.

The paper benchmarks sequencing methods to recover mosaic variants in human brain tissue and introduces RIVER to enhance variant detection across methods.

This study shows that selecting optimal bin sizes improves tumor cell phylogeny inference from low-coverage single-cell sequencing data.