A5026383310- AI in cancer detection
- Bioinformatics and Genomic Networks
- Genetics, Bioinformatics, and Biomedical Research
- Gene expression and cancer classification
- Computational Drug Discovery Methods
- Radiomics and Machine Learning in Medical Imaging
- Artificial Intelligence in Healthcare
- Cancer Genomics and Diagnostics
- Advanced Proteomics Techniques and Applications
- Biomedical Text Mining and Ontologies
- Cell Image Analysis Techniques
- Cancer Diagnosis and Treatment
- Machine Learning in Materials Science
- Gene Regulatory Network Analysis
- Metabolomics and Mass Spectrometry Studies
- SARS-CoV-2 detection and testing
- Pancreatic function and diabetes
- Pluripotent Stem Cells Research
- Congenital heart defects research
- Advanced biosensing and bioanalysis techniques
- Advanced Optical Sensing Technologies
- Optical Systems and Laser Technology
- Ocular and Laser Science Research
- Ubiquitin and proteasome pathways
- Medical Imaging and Pathology Studies
Children's Medical Research Institute
2022-2025
The University of Sydney
2022-2025
Westmead Institute
2022
Nanjing Tech University
2021
Monash University
2014
The proteome provides unique insights into disease biology beyond the genome and transcriptome. A lack of large proteomic datasets has restricted identification new cancer biomarkers. Here, proteomes 949 cell lines across 28 tissue types are analyzed by mass spectrometry. Deploying a workflow to quantify 8,498 proteins, these data capture evidence cell-type post-transcriptional modifications. Integrating multi-omics, drug response, CRISPR-Cas9 gene essentiality screens with deep...
This study introduced a barcode-like design into paper-based blood typing device by integrating with smartphone-based technology. The concept of presenting assay in pattern significantly enhanced the adaptability to smartphone fabrication this involved use printing technique define hydrophilic bar channels which were, respectively, treated Anti-A, -B, and -D antibodies. These were then used perform assays introducing sample. Blood type can be visually identified from eluting lengths...
<p>ROC curves and precision-recall for breast cancer subtype classification</p>
<p>Overview of datasets used in this study</p>
<p>Generalization errors for breast cancer subtype classification</p>
<p>ROC curves and precision-recall for TCGA cancer type classification</p>
<p>Evaluation metrics</p>
<p>Generalization errors for drug response prediction</p>
<p>Consistency of drug response predictions from different machine learning models</p>
<p>Benchmarking results for breast cancer subtype classification with cross-validation</p>
<p>Benchmarking results for drug response prediction with cross-validation</p>
<p>Benchmarking results for cancer type classification with cross-validation</p>
<p>Details of pathway encoder and Transformer encoder.</p>
<p>Ablation study of DeePathNet on cancer type and breast subtype classification</p>
<p>Analysis of performance drug response prediction by target pathways</p>
<p>Computation time for DeePathNet</p>
<p>Ablation study of DeePathNet on drug response prediction</p>
Abstract Multi-omic data analysis incorporating machine learning has the potential to significantly improve cancer diagnosis and prognosis. Traditional methods are usually limited omic measurements, omitting existing domain knowledge, such as biological networks that link molecular entities in various types. Here we develop a Transformer-based explainable deep model, DeePathNet, which integrates cancer-specific pathway information into multi-omic analysis. Using variety of big datasets,...
Abstract Integrating diverse types of biological data is essential for a holistic understanding cancer biology, yet it remains challenging due to heterogeneity, complexity, and sparsity. Addressing this, our study introduces an unsupervised deep learning model, MOSA (Multi-Omic Synthetic Augmentation), specifically designed integrate augment the Cancer Dependency Map (DepMap). Harnessing orthogonal multi-omic information, this model successfully generates molecular phenotypic profiles,...
3120 Background: Accurate tumour classification based upon tissue of origin (TOO) remains important to guide treatment selection and prognosis but can be challenging in patients with poorly differentiated malignancy, cancer unknown primary (CUP) or those prior malignancy. Data-independent acquisition mass spectrometry (DIA-MS)-based proteomics is emerging as a potential clinical diagnostic prognostic tool. We aimed develop protein-based signature identify histological subtype adenocarcinoma...
Abstract Multi-omic data analysis incorporating machine learning has the potential to significantly improve cancer diagnosis and prognosis. Traditional methods are usually limited omic measurements, omitting existing domain knowledge, such as biological networks that link molecular entities in various types. Here we develop a Transformer-based explainable deep model, DeePathNet, which integrates cancer-specific pathway information into multi-omic analysis. Using variety of big datasets,...