Cancer Genomics and Diagnostics

This study uses sequence-based deep learning models to identify functional promoter mutations in cancer, finding enrichment in known and novel cancer genes.

The paper introduces two interpretable neural networks to prioritize cancer driver genes by analyzing genome-wide dependency landscapes and drug-response patterns.

This paper introduces μSeq, a method for accurately quantifying mutation rates using deep sequencing of single clonal expansions, showing it outperforms existing methods in robustness and accuracy.

This study introduces NDRDiff, a computational framework for analyzing cfDNA fragmentation profiles to infer gene expression and identify differentially expressed genes in cancer.

This study uses single-cell multiomics to trace clonal and phenotypic dynamics in hepatoblastoma after cisplatin treatment, identifying a plasticity-proliferation axis driven by BIRC5.

This study develops a pan-cancer PDO platform to preserve tumor heterogeneity and identify therapeutic vulnerabilities, finding that 58% of PDOs respond to Talazoparib and combination screens reveal resistance-overcoming drugs.

This paper proposes a statistical framework using hierarchical mixed-effects models to analyze ctDNA dynamics in liquid biopsy data, finding that a hierarchical linear spline model best captures patient-level ctDNA evolution.

This paper introduces MOBSTERm, a Bayesian framework for subclonal deconvolution that accounts for spatio-temporal sampling biases in cancer sequencing data, improving the analysis of tumour evolution.

This paper introduces TickTack, a Bayesian model to time and cluster co-occurring genome amplifications in cancer, revealing insights into tumor evolution and genomic instability.

This paper describes a protocol for isolating circulating cell-free DNA from murine saliva and validates its use in oral and systemic disease research.

The paper presents a multi-omic resource for Indian breast cancer patients, identifying genetic and proteomic alterations and immune-related gene expression patterns.

This study develops an in vivo platform called SETs to reconstruct tumor evolution without chromosomal instability, identifying therapeutic response determinants in prostate cancer.

This study analyzes 1024 lung adenocarcinomas to identify evolutionary patterns and mutational mechanisms, revealing distinct trajectories linked to smoking, ancestry, and retrotransposition.

SVelfie identifies cancer driver genes by analyzing enrichment of functional structural variants in cancer genomes.

This paper introduces ArtiCull, a method to analyze single-cell whole-genome sequencing data to distinguish active from historical mutational processes in cancer, revealing temporal and spatial patterns in mutational signatures.

SAEG is a deep learning model that predicts gene knockout susceptibility by integrating mutations, copy number alterations, and protein interactions, outperforming existing methods.

This paper introduces Somatic-IV, a method to identify candidate driver genes in cancer by integrating genetic, transcriptomic, and clinical data, and finds ZBED2 and CSNK2A2 as potential drivers in pancreatic and colorectal cancers.

This study identifies somatic repetitive element insertions as new biomarkers in pan-cancer by analyzing 325 cancer genomes using long-read sequencing.

The paper examines genomic diversity within the MDA-MB-231 breast cancer cell line across sublines, identifying intraline heterogeneity driven by single nucleotide variants and copy number alterations.

This study analyzes cancer dependency maps across 1,150 molecular profiles to identify biomarkers for patient stratification and therapeutic response prediction.

SVCFit estimates SV cellular fractions from bulk sequencing to reconstruct clonal evolution in mCRPC under therapy, showing treatment-driven clonal reconfiguration.

The paper shows that hypermutational processes can explain the prevalence of weak driver mutations like atypical KRAS variants in cancers without requiring differential selection.

FiTree infers fitness landscapes from single-cell tumor mutation trees using Bayesian methods, outperforming existing models and identifying epistatic effects in leukemia data.

The paper identifies co-occurring mutations in cancer genomes using FP-Growth, revealing 100,933 mutation pairs across 22 cancer types and highlighting potential therapeutic targets.

The paper analyzes how neighbour proteins influence cancer driver mutations and survival outcomes across multiple cancer types.

This study examines four mouse breast cancer models to show that oncogenic drivers influence metastasis-specific copy number variations and gene expression, while genetic background also plays a significant role.

RNAsum integrates WGS and WTS data to improve cancer diagnostics by corroborating clinically relevant variants and supporting therapeutic decisions.

The paper analyzes somatic mutations across tissues and individuals using exome sequencing data from GTEx donors, identifying 8,470 variants and providing insights into mutation rates and mechanisms.

This study compares deep learning and machine learning for biomarker prediction in cancer genomes, finding DL performs better in cross-assay and data-sparse settings.

PHALCON improves variant calling and tumor phylogeny reconstruction in large-scale single-cell panel sequencing data by modeling tumor evolution under a finite-sites model.