Adel T. Aref

Adel T. Aref

Auto-generated profile, based on publicly available information

Adel T. Aref

ProCan, Children’s Medical Research Institute, Faculty of Medicine and Health, The University of Sydney, Westmead, Australia

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 December 2024 in Cancer Research Communications
Verified
Journal Article
Open access
ZCZhaoxiang Cai
Zhaoxiang Cai
RPRebecca Poulos
Rebecca Poulos
AAAdel Aref
Adel Aref
PRPhillip J. Robinson
Phillip J. Robinson
RRRoger Reddel
Roger Reddel
QZQing Zhong
Qing Zhong
Multiomic 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 multiomic data analysis. Using a variety of big datasets, including ProCan-DepMapSanger, Cancer Cell Line Encyclopedia, and The Cancer Genome Atlas, 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. Significance: DeePathNet integrates cancer-specific biological pathways using transformer-based deep learning for enhanced cancer analysis. It outperforms existing models in predicting drug responses, cancer types, and subtypes. By enabling pathway-level biomarker discovery, DeePathNet represents a significant advancement in cancer research and could lead to more effective treatments.

4 December 2023 in Life Science Alliance
Verified
Journal Article
Open access
PDF file
QZQing Zhong
Qing Zhong
RSRui Sun
Rui Sun
AAAdel T Aref
Adel T Aref
ZNZainab Noor
Zainab Noor
AAAsim Anees
Asim Anees
PWPeter Wild
Peter Wild
Gleason grading is an important prognostic indicator for prostate adenocarcinoma and is crucial for patient treatment decisions. However, intermediate-risk patients diagnosed in the Gleason grade group (GG) 2 and GG3 can harbour either aggressive or non-aggressive disease, resulting in under- or overtreatment 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 dichotomize 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.

9 June 2025 in Cancer Discovery
Verified
Journal Article
Open access
ZCZhaoxiang Cai
Zhaoxiang Cai
EBEmma Boys
Emma Boys
ZNZainab Noor
Zainab Noor
AAAdel T. Aref
Adel T. Aref
DXDylan Xavier
Dylan Xavier
RRRoger Reddel
Roger Reddel
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 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 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 with local models and matching centralized model performance. The approach’s generalizability was demonstrated by retraining the global model with data from two external, data-independent acquisition mass spectrometry cohorts (n = 55) and eight acquired by tandem mass tag proteomics (n = 832). ProCanFDL presents a solution for internationally collaborative machine learning initiatives using proteomic data, for example, for discovering predictive biomarkers or treatment targets while maintaining data privacy. Significance: A federated deep learning approach applied to human proteomic data, acquired using two distinct proteomic technologies from 40 tumor cohorts across eight countries, enabled accurate cancer histopathologic subtyping while preserving data privacy. This approach will enable the privacy-compliant development of large-scale proteomic artificial intelligence models, including foundation models, across institutions globally.

22 December 2025 in Clinical Cancer Research
Verified
Journal Article
Open access
NMNicholas J. McNamee
Nicholas J. McNamee
JLJia Liu
Jia Liu
RPRebecca Poulos
Rebecca Poulos
AAAdel T. Aref
Adel T. Aref
RRRoger Reddel
Roger Reddel
Antibody–drug conjugate (ADC) development is progressing rapidly, with significant impacts on the treatment landscape of clinical oncology. However, the development of predictive biomarkers to guide ADC selection has not kept pace and to date has been mostly limited to immunohistochemistry, hampering patient selection and personalized treatment recommendations. Here, we review the current state of ADC target protein assessment and the association between target expression levels and therapeutic efficacy. We discuss the limitations and challenges of existing protein assessment technology and highlight the potential clinical role of quantitative mass spectrometry–based proteomics for rapid, comprehensive, and accurate protein quantification. We propose that integrating multiplexed proteomic assays early in ADC development and clinical trial design can improve therapeutic targeting, enhance clinical trial design, and ultimately lead to more personalized ADC-based treatment strategies.

31 October 2022 in bioRxiv
Verified
Preprint
Open access
PDF file
ZCZhaoxiang Cai
Zhaoxiang Cai
RPRebecca Poulos
Rebecca Poulos
AAAdel Aref
Adel Aref
PRPhillip J. Robinson
Phillip J. Robinson
RRRoger Reddel
Roger Reddel
QZQing Zhong
Qing Zhong
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 . Highlights: DeePathNet integrates biological pathways for enhanced cancer analysis. DeePathNet utilizes Transformer-based deep learning for superior accuracy. DeePathNet outperforms existing models in drug response prediction. DeePathNet enables pathway-level biomarker discovery in cancer research.

1 October 2025 in Cancer Research Communications
Verified
Journal Article
Open access
AAAdel T Aref
Adel T Aref
JGJason Grealey
Jason Grealey
MPMohashin Pathan
Mohashin Pathan
ZNZainab Noor
Zainab Noor
AAAsim Anees
Asim Anees
AGAnthony J Gill
Anthony J Gill
Pancreatic ductal adenocarcinoma (PDA) is an aggressive malignancy that lacks reliable biomarkers to guide treatment decisions. Effective prognostic tools are needed to improve its clinical management. We conducted a comprehensive proteomic analysis on 115 PDA patient samples with matched adjacent normal tissue. A 20-protein diagnostic panel was identified (LGALS1, ANXA2, LGALS3BP, CTSD, S100P, COL12A1, SFN, THBS2, CTHRC1, THBS1, SERPINB5, LAMC2, POSTN, CEACAM6, CTSE, PLEC, PKM, S100A11, TAGLN2, ALDOA). Consensus clustering analysis identified four prognostic proteomic subtypes. Subtypes with poorer prognoses exhibited upregulation of neutrophil degranulation, extracellular matrix remodeling, focal adhesion, Mesenchymal Epithelial Transition, collagen formation, and PI3K-Akt-mTOR-related pathways, indicating a predominance of basal-like and activated stromal features. In tumors with homologous recombination deficiency or Catalogue of Somatic Mutations in Cancer Signature-3, several immune-related proteins were enriched. An 18-protein (PURB, SDCBP2, CD2BP2, GALM, SERPINA3, OAS3, FAN1, ZPR1, KRT2, NUDT2, SMNDC1, SERPINA4, CUTA, WDR36, POSTN, CLEC11A, PEX14, and PI4KA) risk score was developed and validated using multicox regression analyses with LASSO regularization. The risk score demonstrated independent prognostic significance for overall survival and recurrence, and was validated in an independent proteomic dataset generated using a different proteomic technology. This study thus introduces four novel prognostic PDA subtypes, and an 18-protein risk score validated in an independent dataset, which shows promise for improving survival prediction and could serve as a valuable tool for personalized treatment guidance. Significance: The findings from this study have significant implications for the future of pancreatic cancer management. By identifying a 20-protein panel with diagnostic and screening potential, this research provides a foundation for developing early detection tools for PDA, an aggressive cancer with limited treatment options. The classification of PDA into four proteomic subtypes with distinct prognostic outcomes paves the way for subtype-specific therapeutic approaches, allowing clinicians to better stratify patients based on their risk profiles. Additionally, the validated 18-protein risk score, which enhances survival prediction and operates independently of existing clinical variables, represents a promising tool for personalized prognostic assessments. Incorporating these proteomic-based biomarkers into clinical practice could improve diagnostic accuracy, guide individualized treatment decisions, and ultimately enhance patient outcomes in PDA. This study underscores the potential of proteomic profiling to improve cancer treatment by providing targeted, actionable insights into tumor biology.