Biology subjects

Poulos, R. C.

Publications and source records attributed to Poulos, R. C..

9 recordsLinked to original sources

bioRxiv2025

OnCorr: A pan-cancer mRNA-protein correlation tool for precision oncology

Proteins are ultimately responsible for cellular phenotypes and are targeted by most anticancer drugs. However, beyond immunohistochemistry, proteins are not typically measured in precision oncology, meaning transcriptomics is used as a proxy. To determine how informative mRNA is for guiding personalised treatments, mRNA-protein correlations were analysed in three large pan-cancer datasets and made available in a web portal (https://procan.shinyapps.io/OnCorr/). OnCorr can be integrated into precision medicine programs to augment transcriptomics.

cancer biology↗

bioRxiv2025

Potent synthetic lethality between PLK1 and EYA-family inhibitors in tumours of the central and peripheral nervous system

The Eyes Absent family of protein phosphatases (EYA1-4) are aberrantly expressed and tumour-promoting across many devastating cancers of neurological origin affecting both children and adults. It has recently been demonstrated that EYA1 and EYA4 promote tumour cell survival by increasing the active pool of Polo-like kinase 1 (PLK1) molecules. This discovery provides a rationale for the therapeutic combination of EYA inhibitors with direct, ATP-competitive, PLK1 inhibitors. Here, we demonstrate potent and synergistic effects of EYA and PLK1 inhibition in cancer cell lines that overexpress EYA1 and/or EYA4, including in neuroblastoma and glioblastoma models. We identify decreases in PLK1 activity and RAD51 foci formation, and increases in mitotic arrest and cell death, as mechanistic contributors to combination sensitivity. Combined EYA and PLK1 inhibition is also effective in glioblastoma stem cell models that overexpress EYA1/EYA4 and specifically targets the cancer stem cell state. Finally, through multi-omic correlational analysis, we identify high levels of the NuRD complex and SOX9 as contributors to combination treatment sensitivity. Overall, this work identifies a novel synthetic lethal combination therapy with potential utility across a wide range of neurological cancers.

cancer biology↗

bioRxiv2024

Federated deep learning enables cancer subtyping by proteomics

Artificial intelligence applications in biomedicine face major challenges from data privacy requirements. To address this issue for clinically annotated tissue proteomic data, we developed a Federated Deep Learning (FDL) approach (ProCanFDL), training local models on simulated sites containing data from a pan-cancer cohort (n=1,260) and 29 cohorts held behind private firewalls (n=6,265), representing 19,930 replicate data-independent acquisition mass spectrometry (DIA-MS) runs. Local parameter updates were aggregated to build the global model, achieving a 43% performance gain on the hold-out test set (n=625) in 14 cancer subtyping tasks compared to local models, and matching centralized model performance. The approachs generalizability was demonstrated by retraining the global model with data from two external DIA-MS cohorts (n=55) and eight acquired by tandem mass tag (TMT) proteomics (n=832). ProCanFDL presents a solution for internationally collaborative machine learning initiatives using proteomic data, e.g., for discovering predictive biomarkers or treatment targets, while maintaining data privacy. Statement of SignificanceA federated deep learning approach applied to human proteomic data, acquired using two distinct proteomic technologies from 40 tumor cohorts from eight countries, enabled accurate cancer histopathological subtyping while preserving data privacy. This approach will enable privacy-compliant development of large-scale proteomic AI models, including foundation models, across institutions globally.

cancer biology↗

bioRxiv2023

Proteomic-based stratification of intermediate-risk prostate cancer patients

Gleason grading is an important prognostic indicator for prostate adenocarcinoma and is crucial for patient treatment decisions. However, intermediate-risk patients diagnosed in Gleason Grade Groups (GG) 2 and GG3 can harbour either aggressive or non-aggressive disease, resulting in under- or over-treatment of a significant number of patients. Here, we performed proteomic, differential expression, machine learning, and survival analyses for 1,348 matched tumour and benign sample runs from 278 patients. Three proteins (F5, TMEM126B and EARS2) were identified as candidate biomarkers in patients with biochemical recurrence. Multivariate Cox regression yielded 18 proteins, from which a risk score was constructed to dichotomise prostate cancer patients into low- and high-risk groups. This 18-protein signature is prognostic for the risk of biochemical recurrence and completely independent of the intermediate GG. Our results suggest that markers generated by computational proteomic profiling have the potential for clinical applications including integration into prostate cancer management.

cancer biology↗

bioRxiv2022

Transformer-based deep learning integrates multi-omic data with cancer pathways

Multi-omic data analysis incorporating machine learning has the potential to significantly improve cancer diagnosis and prognosis. Traditional machine learning methods are usually limited to omic measurements, omitting existing domain knowledge, such as the biological networks that link molecular entities in various omic data types. Here we develop a Transformer-based explainable deep learning model, DeePathNet, which integrates cancer-specific pathway information into multi-omic data analysis. Using a variety of big datasets, including ProCan-DepMapSanger, CCLE, and TCGA, we demonstrate and validate that DeePathNet outperforms traditional methods for predicting drug response and classifying cancer type and subtype. Combining biomedical knowledge and state-of-the-art deep learning methods, DeePathNet enables biomarker discovery at the pathway level, maximizing the power of data-driven approaches to cancer research. DeePathNet is available on GitHub at https://github.com/CMRI-ProCan/DeePathNet. HighlightsO_LIDeePathNet integrates biological pathways for enhanced cancer analysis. C_LIO_LIDeePathNet utilizes Transformer-based deep learning for superior accuracy. C_LIO_LIDeePathNet outperforms existing models in drug response prediction. C_LIO_LIDeePathNet enables pathway-level biomarker discovery in cancer research. C_LI

bioinformatics↗

bioRxiv2022

Pan-cancer proteomic map of 949 human cell lines reveals principles of cancer vulnerabilities

The proteome provides unique insights into biology and disease beyond the genome and transcriptome. Lack of large proteomic datasets has restricted identification of new cancer biomarkers. Here, proteomes of 949 cancer cell lines across 28 tissue types were analyzed by mass spectrometry. Deploying a clinically-relevant workflow to quantify 8,498 proteins, these data capture evidence of cell type and post-transcriptional modifications. Integrating multi-omics, drug response and CRISPR-Cas9 gene essentiality screens with a deep learning-based pipeline revealed thousands of protein-specific biomarkers of cancer vulnerabilities. Proteomic data had greater power to predict drug response than the equivalent portion of the transcriptome. Further, random downsampling to only 1,500 proteins had limited impact on predictive power, consistent with protein networks being highly connected and co-regulated. This pan-cancer proteomic map (ProCan-DepMapSanger), available at https://cellmodelpassports.sanger.ac.uk, is a comprehensive resource revealing principles of protein regulation with important implications for future clinical studies.

systems biology↗

bioRxiv2019

Mutational processes of distinct POLE exonuclease domain mutants drive an enrichment of a specific TP53 mutation in colorectal cancer

Cancer genomes with mutations in the exonuclease domain of Polymerase Epsilon (POLE) present with an extraordinarily high somatic mutation burden. In vitro studies have shown that distinct POLE mutants exhibit different polymerase activity and yet, how these POLE mutants generate mutations across cancer genomes and influence driver events remains poorly understood. Here we analyzed 7,345 colorectal cancer samples, including nine whole genome sequenced samples harboring POLE mutations. Our analysis identified differential mutation spectra across the mutants including methylation-independent enrichment of C>T mutations in POLE V411L. In contrast, analysis of other genomic regions showed similar mutation profiles across the different POLE mutants. Notably, we found that POLE mutants with the TP53 R213* mutation, caused by a TT[C>T]GA substitution, have significantly higher relative frequency of this mutational context compared with samples without this mutation. This finding demonstrates that variations in underlying mutation spectra can increase the likelihood of specific driver mutation formation.

genomics↗

bioRxiv2018

In search of non-coding driver mutations by deep sequencing of regulatory elements in colorectal cancer

Large-scale whole cancer-genome sequencing projects have led to the identification of a handful of cis-regulatory driver mutations in cancer genomes. However, recent studies have demonstrated that very large cancer cohorts will be required in order to identify low frequency non-coding drivers. To further this endeavour, in this study, we performed highdepth sequencing across 95 colorectal cancers and matched normal samples using a unique target capture sequencing (TCS) assay focusing on over 35 megabases of gene regulatory elements. We first assessed coverage and variant detection capability from our TCS data, and compared this with a sample that was additionally whole-genome sequenced (WGS). TCS enabled substantially deeper sequencing and thus we detected 51% more somatic single nucleotide variants (n = 2,457) and 144% more somatic insertions and deletions (n = 39) by TCS than WGS. Variants obtained from TCS data were suitable for somatic mutational signature detection, enabling us to define the signatures associated with germline deleterious variants in MSH6 and MUTYH in samples within our cohort. Finally, we surveyed regulatory mutations to find putative drivers by assessing variant recurrence and function, identifying some regulatory variants that may influence oncogenesis. Our study demonstrates TCS to be a sequencing-efficient alternative to traditional WGS, enabling improved coverage and variant detection when seeking to identify variants at specific loci among larger cohorts. Interestingly, we found no candidate variants that have a clear driver function, suggesting that regulatory drivers may be rare in a colorectal cancer cohort of this size.\n\nAuthor SummaryIn recent years, some cancer research focus has turned towards the role of somatic mutations in the 98% of the genome that is non-coding. To investigate such mutations, we performed deep sequencing of regulatory regions and a selection of coding genes across 95 colorectal cancer and matched-normal samples. To determine the ability of our targeted deep sequencing methodology to accurately detect variants, we compared our results with those from a sample that was additionally whole-genome sequenced. We found target capture sequencing to enable greater sequencing depth, allowing the detection of 51% and 144% more somatic single nucleotide and insertion/deletion mutations, respectively. Our study here demonstrates target capture sequencing to be a useful approach for researchers seeking to identify variants at specific loci among larger cohorts. Our results also enabled the generation of mutational signatures, implicating deleterious germline single nucleotide variants in coding exons of MSH6 and MUTYH in samples within our cohort. Finally, we surveyed regulatory elements in search of somatic cancer driver mutations. We identified some regulatory variants that may influence oncogenesis, but found no candidate variants with clear driver function. These findings suggest that regulatory driver mutations may be rare in a colorectal cancer cohort of this size.

biochemistry↗

bioRxiv2017

CpG methylation accounts for genome-wide C>T mutation variation and cancer driver formation across cancer types

Cytosine methylation (5mC) is vital for cellular function, and yet 5mC sites are also commonly mutated in the genome. In this study, we analyse the genomes of over 900 cancer samples, together with tissue type-specific methylation and replication timing data. We describe a strong mutation-methylation association in colorectal cancers with microsatellite instability (MSI) or with Polymerase epsilon (POLE) exonuclease domain mutation. We describe a potential role for mismatch repair in the correction of mismatches resulting from deamination of 5mC, and propose a mutator phenotype to exist in POLE-mutant cancers specifically at 5mC sites. We also associate POLE-mutant hotspot coding mutations in APC and TP53 with CpG methylation. Analysing mutations across additional cancer types, we identify nucleotide excision repair- and AID/APOBEC-induced processes to underlie differential mutation-methylation associations in certain cancer subtypes. This study reveals differential associations vital for accurately mapping regional variation in mutation density and pinpointing driver mutations in cancer.

genomics↗