Model ensembling as a tool to form interpretable multi-omic predictors of cancer pharmacosensitivity DOI Creative Commons
Sébastien De Landtsheer, Apurva Badkas, Dagmar Kulms

и другие.

Briefings in Bioinformatics, Год журнала: 2024, Номер 25(6)

Опубликована: Сен. 23, 2024

Abstract Stratification of patients diagnosed with cancer has become a major goal in personalized oncology. One important aspect is the accurate prediction response to various drugs. It expected that molecular characteristics cells contain enough information retrieve specific signatures, allowing for predictions based solely on these multi-omic data. Ideally, should be explainable clinicians, order integrated care. We propose machine-learning framework ensemble learning integrate data and predict sensitivity an array commonly used experimental compounds, including chemotoxic compounds targeted kinase inhibitors. trained set classifiers different parts our dataset produce omic-specific then random forest classifier signatures drug responsiveness. Cancer Cell Line Encyclopedia dataset, comprising measurements hundreds cell lines, build predictive models, validated results using nested cross-validation. Our show good performance several (Area under Receiver-Operating Curve >79%) across most frequent types. Furthermore, simplicity approach allows examine which omic layers have greater importance models identify new putative markers small subsets transcriptional potential useful tools oncology, paving way clinicians use tumors therapeutic compounds.

Язык: Английский

TMO-Net: an explainable pretrained multi-omics model for multi-task learning in oncology DOI Creative Commons

Feng-ao Wang,

Zhenfeng Zhuang,

Feng Gao

и другие.

Genome biology, Год журнала: 2024, Номер 25(1)

Опубликована: Июнь 6, 2024

Abstract Cancer is a complex disease composing systemic alterations in multiple scales. In this study, we develop the Tumor Multi-Omics pre-trained Network (TMO-Net) that integrates multi-omics pan-cancer datasets for model pre-training, facilitating cross-omics interactions and enabling joint representation learning incomplete omics inference. This enhances sample empowers various downstream oncology tasks with datasets. By employing interpretable learning, characterize contributions of distinct features to clinical outcomes. The TMO-Net serves as versatile framework cross-modal oncology, paving way tumor omics-specific foundation models.

Язык: Английский

Процитировано

9
Artificial intelligence applied to ‘omics data in liver disease: towards a personalised approach for diagnosis, prognosis and treatment DOI Creative Commons
Soumita Ghosh, Xun Zhao,

Mouaid Alim

и другие.

Gut, Год журнала: 2024, Номер unknown, С. gutjnl - 331740

Опубликована: Авг. 22, 2024

Advancements in omics technologies and artificial intelligence (AI) methodologies are fuelling our progress towards personalised diagnosis, prognosis treatment strategies hepatology. This review provides a comprehensive overview of the current landscape AI methods used for analysis data liver diseases. We present an prevalence different levels across various diseases, as well categorise methodology studies. Specifically, we highlight predominance transcriptomic genomic profiling relatively sparse exploration other such proteome methylome, which represent untapped potential novel insights. Publicly available database initiatives The Cancer Genome Atlas International Consortium have paved way advancements diagnosis hepatocellular carcinoma. However, same availability large datasets remains limited Furthermore, application sophisticated to handle complexities multiomics requires substantial train validate models faces challenges achieving bias-free results with clinical utility. Strategies address paucity capitalise on opportunities discussed. Given global burden chronic it is imperative that multicentre collaborations be established generate large-scale early disease recognition intervention. Exploring advanced also necessary maximise these improve detection strategies.

Язык: Английский

Процитировано

9
Artificial intelligence for medicine 2025: Navigating the endless frontier DOI
Jiyan Dai, Huiyu Xu, Tao Chen

и другие.

The Innovation Medicine, Год журнала: 2025, Номер unknown, С. 100120 - 100120

Опубликована: Янв. 1, 2025

<p>Artificial intelligence (AI) is driving transformative changes in the field of medicine, with its successful application relying on accurate data and rigorous quality standards. By integrating clinical information, pathology, medical imaging, physiological signals, omics data, AI significantly enhances precision research into disease mechanisms patient prognoses. technologies also demonstrate exceptional potential drug development, surgical automation, brain-computer interface (BCI) research. Through simulation biological systems prediction intervention outcomes, enables researchers to rapidly translate innovations practical applications. While challenges such as computational demands, software ethical considerations persist, future remains highly promising. plays a pivotal role addressing societal issues like low birth rates aging populations. can contribute mitigating rate through enhanced ovarian reserve evaluation, menopause forecasting, optimization Assisted Reproductive Technologies (ART), sperm analysis selection, endometrial receptivity fertility remote consultations. In posed by an population, facilitate development dementia models, cognitive health monitoring strategies, early screening systems, AI-driven telemedicine platforms, intelligent smart companion robots, environments for aging-in-place. profoundly shapes medicine.</p>

Язык: Английский

Процитировано

1
Contrastive Learning for Omics-guided Whole-slide Visual Embedding Representation DOI Creative Commons

Suwan Yu,

Yooeun Kim, Hyun-Seok Kim

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2025, Номер unknown

Опубликована: Янв. 15, 2025

Abstract While computational pathology has transformed cancer diagnosis and prognosis prediction, existing methods remain limited in their ability to decipher the complex molecular characteristics within tumors. We present CLOVER (Contrastive Learning for Omics-guided whole-slide Visual Embedding Representation), a novel deep learning framework that leverages self-supervised contrastive integrate multi-omics data (genomics, epigenomics, transcriptomics) with slide representations, connecting morphological features of Using breast cohorts comprising diagnostic slides paired from 610 patients, we validated CLOVER’s excellence by demonstrating its generate effective slide-level representations consider states cancer. outperforms few-shot scenarios, particularly subtype classification clinical biomarker prediction tasks (ER, PR, HER2 status). Through comprehensive interpretability analysis, identified tumor microenvironment components revealed associated Our results demonstrate enables detailed characterization single suggesting potential utilization future studies.

Язык: Английский

Процитировано

0
Enhancing Molecular Network‐Based Cancer Driver Gene Prediction Using Machine Learning Approaches: Current Challenges and Opportunities DOI Creative Commons
Hao Zhang,

Chaohuan Lin,

Y. Chen

и другие.

Journal of Cellular and Molecular Medicine, Год журнала: 2025, Номер 29(1)

Опубликована: Янв. 1, 2025

ABSTRACT Cancer is a complex disease driven by mutations in the genes that play critical roles cellular processes. The identification of cancer driver crucial for understanding tumorigenesis, developing targeted therapies and identifying rational drug targets. Experimental validation are time‐consuming costly. Studies have demonstrated interactions among associated with similar phenotypes. Therefore, using molecular network‐based approaches necessary. Molecular random walk‐based approaches, which integrate mutation data protein–protein interaction networks, been widely employed predicting robust predictive potential. However, recent advancements deep learning, particularly graph‐based models, provided novel opportunities enhancing prediction genes. This review aimed to comprehensively explore how machine learning methodologies, network propagation, graph neural autoencoders, embeddings, attention mechanisms, improve scalability interpretability gene prediction.

Язык: Английский

Процитировано

0
Pan‐cancer analysis shapes the understanding of cancer biology and medicine DOI Creative Commons
Xiaoping Cen,

Yuanyuan Lan,

Jiansheng Zou

и другие.

Cancer Communications, Год журнала: 2025, Номер unknown

Опубликована: Март 22, 2025

Abstract Advances in multi‐omics datasets and analytical methods have revolutionized cancer research, offering a comprehensive, pan‐cancer perspective. Pan‐cancer studies identify shared mechanisms unique traits across different types, which are reshaping diagnostic treatment strategies. However, continued innovation is required to refine these approaches deepen our understanding of biology medicine. This review summarized key findings from research explored their potential drive future advancements oncology.

Язык: Английский

Процитировано

0
PCLSurv: a prototypical contrastive learning-based multi-omics data integration model for cancer survival prediction DOI Creative Commons
Zhimin Li, Wenlan Chen, Hai Zhong

и другие.

Briefings in Bioinformatics, Год журнала: 2025, Номер 26(2)

Опубликована: Март 1, 2025

Accurate cancer survival prediction remains a critical challenge in clinical oncology, largely due to the complex and multi-omics nature of data. Existing methods often struggle capture comprehensive range informative features required for precise predictions. Here, we introduce PCLSurv, an innovative deep learning framework designed using PCLSurv integrates autoencoders extract omics-specific employs sample-level contrastive identify distinct yet complementary characteristics across data views. Then, are fused via bilinear fusion module construct unified representation. To further enhance model's capacity high-level semantic relationships, aligns similar samples with shared prototypes while separating unrelated ones prototypical learning. As result, effectively distinguishes patient groups varying outcomes at different similarity levels, providing robust stratifying patients based on molecular features. We conduct extensive experiments 11 datasets. The comparison results confirm superior performance over existing alternatives. source code is freely available https://github.com/LiangSDNULab/PCLSurv.

Язык: Английский

Процитировано

0
Deep Learning of radiology-genomics integration for computational oncology: A mini review DOI Creative Commons

Feng-ao Wang,

Y Li, Tao Zeng

и другие.

Computational and Structural Biotechnology Journal, Год журнала: 2024, Номер 23, С. 2708 - 2716

Опубликована: Июнь 20, 2024

In the field of computational oncology, patient status is often assessed using radiology-genomics, which includes two key technologies and data, such as radiology genomics. Recent advances in deep learning have facilitated integration radiology-genomics even new omics significantly improving robustness accuracy clinical predictions. These factors are driving artificial intelligence (AI) closer to practical applications. particular, models crucial identifying biomarkers therapeutic targets, supported by explainable AI (xAI) methods. This review focuses on recent developments for integration, highlights current challenges, outlines some research directions multimodal biomarker discovery or radiology-omics that urgently needed oncology.

Язык: Английский

Процитировано

2
Model ensembling as a tool to form interpretable multi-omic predictors of cancer pharmacosensitivity DOI Creative Commons
Sébastien De Landtsheer, Apurva Badkas, Dagmar Kulms

и другие.

Briefings in Bioinformatics, Год журнала: 2024, Номер 25(6)

Опубликована: Сен. 23, 2024

Abstract Stratification of patients diagnosed with cancer has become a major goal in personalized oncology. One important aspect is the accurate prediction response to various drugs. It expected that molecular characteristics cells contain enough information retrieve specific signatures, allowing for predictions based solely on these multi-omic data. Ideally, should be explainable clinicians, order integrated care. We propose machine-learning framework ensemble learning integrate data and predict sensitivity an array commonly used experimental compounds, including chemotoxic compounds targeted kinase inhibitors. trained set classifiers different parts our dataset produce omic-specific then random forest classifier signatures drug responsiveness. Cancer Cell Line Encyclopedia dataset, comprising measurements hundreds cell lines, build predictive models, validated results using nested cross-validation. Our show good performance several (Area under Receiver-Operating Curve &gt;79%) across most frequent types. Furthermore, simplicity approach allows examine which omic layers have greater importance models identify new putative markers small subsets transcriptional potential useful tools oncology, paving way clinicians use tumors therapeutic compounds.

Язык: Английский

Процитировано

0