Trevor A. Graham

This study develops a neuroblastoma mouse model to investigate chemotherapy resistance and immune evasion mechanisms, identifying CDK8 upregulation and immune cell infiltration as key factors.

This study examines the epigenetic evolution of colorectal cancer and its microenvironment, identifying new therapeutic targets through single-cell analysis and functional validation.

Classpose improves cell segmentation and phenotyping in colorectal cancer WSIs, identifying novel spatial morphological phenotypes linked to molecular conditions.

This study identifies PKN2 as a regulator of tight junctions that limits colitis and colorectal cancer by maintaining intestinal barrier integrity during inflammation.

This study uses whole genome sequencing of endometrial cancers to identify novel subgroups, drivers, and actionable alterations, improving precision medicine approaches.

This paper introduces SCOUT, a synthetic whole-genome sequencing resource for benchmarking cancer genomic analysis, revealing performance issues in complex tumours and highlighting the importance of evolutionary ground truth.

This study uses spatial transcriptomic and proteomic profiling to identify cell interaction patterns that predict progression of Barrett's esophagus.

PHYFUM reconstructs the evolutionary history of normal and pre-malignant tissues using methylation data, revealing early gland divergence in the gut and later divergence in the endometrium.

This paper identifies cross resistance between cancer drugs using single-drug exposure data, showing that high cross resistance reduces the effectiveness of switching between CDK4/6 inhibitors.

The paper tracks DNA methylation changes in blood samples up to 15 years before AML diagnosis, identifying DMRs linked to clonal expansions and epigenetic reprogramming.

This paper develops a mathematical framework to infer cancer drug resistance dynamics from genetic lineage tracing and population size data, validated using experimental evolution and single-cell sequencing.

This paper introduces dATAC, a single-cell multi-omic method for simultaneous whole genome sequencing and chromatin accessibility profiling, revealing genome-epigenome coevolution in colorectal cancer.

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.

This paper identifies long deletion signatures in repetitive regions as markers of microsatellite instability and develops a classifier to detect MSI in low-quality sequencing data.

This study shows that immune evasion in colorectal cancer follows a 'Big Bang' pattern, where early genetic and epigenetic changes drive subsequent tumor-immune evolution.

CNETML infers phylogenies from spatio-temporal copy number profiles of cancer samples, handling total copy numbers and temporal data to estimate mutation rates and tree topology.

This paper identifies mutational signatures caused by formalin fixation and introduces FFPEsig, a method to correct these artefacts in genome sequencing data.

The paper introduces rapidly oscillating molecular clocks to quantify human stem cell dynamics and finds differences in stem cell replacement rates between small intestine and colon, as well as effects of germline mutations.

LiquidCNA tracks subclonal evolution from liquid biopsies using somatic copy number alterations, revealing new subpopulations in metastatic lung cancer.

This paper introduces a model-based subclonal reconstruction method that integrates evolutionary theory with machine learning to improve accuracy in tumor subclonal analysis.

The paper quantifies tumor immunogenicity using dN/dS ratios in the immunopeptidome and shows it predicts immunotherapy response by distinguishing adaptation from escape mechanisms.

This study links dN/dS ratios to selective coefficients in somatic evolution, revealing the distribution of fitness effects for mutations in human tissues.

This paper shows that spatial tumour structure affects mutation patterns in sequencing data and introduces a framework to account for these effects in cancer genomic analysis.

NeoPredPipe predicts and assesses neoantigen recognition potential from tumor variants in single and multi-region samples, providing insights into neoantigen burden and tumor heterogeneity.

This paper introduces a machine learning method to detect repeated cancer evolution patterns across patients using multi-region sequencing data, identifying prognostic patient subgroups.

The paper responds to criticism of a method used to infer neutral cancer evolution from variant allele frequency distributions, arguing that neutral evolution explains intra-tumour heterogeneity in some cancers.