With the advancement of high-throughput biotechnologies, we increasingly accumulate biomedical data about diseases, especially cancer. There is a need for computational models and methods to sift through, integrate, and extract new knowledge from the diverse available data to improve the mechanistic understanding of diseases and patient care. To uncover molecular mechanisms and drug indications for specific cancer types, we develop an integrative framework able to harness a wide range of diverse molecular and pan-cancer data. We show that our approach outperforms competing methods and can identify new associations. Furthermore, through the joint integration of data sources, our framework can also uncover links between cancer types and molecular entities for which no prior knowledge is available. Our new framework is flexible and can be easily reformulated to study any biomedical problems.
Integrative Adaptive Indexes from Noisy Routine Haematological Markers can Predict and Discriminate Health Status and Biological Age
Oxford Immune Algorithmics（牛津免疫算法公司）；Oxford University Innovation（牛津大学创新）；Department of Surgery, University of Cambridge（剑桥大学外科系）；British Society for Research on Ageing（英国衰老研究学会）；King’s Institute for Artificial Intelligence, King’s College London（伦敦国王学院人工智能研究所）；Algorithmic Dynamics Lab, Research Departments of Biomedical Computing and Digital Twins, School of Biomedical Engineering and Imaging Sciences, King’s College London（伦敦国王学院生物医学工程与成像科学学院生物医学计算与数字孪生研究部算法动力学实验室）
Pan-Cancer Integrative Histology-Genomic Analysis via Interpretable Multimodal Deep Learning
Brigham and Women’s Hospital（布莱根妇女医院）；Harvard Medical School（哈佛医学院）；Broad Institute of Harvard and MIT（哈佛-麻省理工学院博德研究所）；Dana-Farber Cancer Institute（丹娜-法伯癌症研究所）；Harvard University（哈佛大学）
Machine Learning for Integrating Data in Biology and Medicine: Principles, Practice, and Opportunities
Stanford University（斯坦福大学）；University of Toronto（多伦多大学）；Princess Margaret Cancer Centre（玛格丽特公主癌症中心）；Hikvision Research Institute（海康威视研究院）；Chan Zuckerberg Biohub（陈·扎克伯格生物中心）；SickKids Research Institute（病童医院研究所）；Vector Institute（向量研究所）
Subtype-Former: a deep learning approach for cancer subtype discovery with multi-omics data
East China University of Science and Technology（华东理工大学）；Jacobs Technion-Cornell Institute（雅各布布·泰克尼翁-康奈尔研究所）；Cornell Tech（康奈尔科技校区）；Cornell University（康奈尔大学）
A Hybrid Machine Learning Model for Classifying Gene Mutations in Cancer using LSTM, BiLSTM, CNN, GRU, and GloVe
Maharishi International University（玛赫西国际大学）；Luther College（路德学院）；Al-Balqa Applied University（巴尔卡应用大学）；Skyline University College（天际线大学学院）；Sultan Qaboos University（苏丹卡布斯大学）
Comparisons of Graph Neural Networks on Cancer Classification Leveraging a Joint of Phenotypic and Genetic Features
Mayo Clinic（梅奥诊所）
癌症每年在全世界造成数百万人死亡。尽管癌症医学已经取得了显著的进展，但仍有许多问题有待解决，适当的癌症患者分层是选择适当治疗方案的前提，因为已知癌症患者具有异质性的遗传组成和表型差异。在这项研究中，建立在从梅奥诊所电子健康记录（EHR）和癌症患者遗传检测报告中提取的深度表型特征的基础上，我们评估了各种图神经网络（GNN），这些神经网络利用表型和遗传特征的联合进行癌症类型分类。在Mayo Clinic癌症疾病数据集上应用并微调模型。通过报告的准确度、精确度、召回率和F1值以及基于疾病类别的F1评分进行评估。根据我们的评估结果，GNN的平均表现优于基线模型，平均统计数据始终高于基线模型（准确率为0.849 vs 0.772，精确率为0.858 vs 0.794，召回率为0.843 vs 0.759，F1得分为0.843 vs 0.855）。在GNN中，ChebNet，GraphSAGE和TAGCN表现最好，而GAT表现最差。我们在Mayo Clinic癌症疾病数据集上应用并比较了八个GNN模型，包括AGNN，ChebNet，GAT，GCN，GIN，GraphSAGE，SGC和TAGCN，并评估了它们的性能，并将它们相互比较，并与我们用作基线的更传统的机器学习模型，如决策树，梯度提升，多层感知器，朴素贝叶斯和随机森林进行比较。
Federated unsupervised random forest for privacy-preserving patient stratification
Institute for Medical Informatics, Statistics and Documentation Medical University Graz（格拉茨医科大学医学信息学、统计学与文献学研究所）；Department of Pure and Applied Sciences, University of Urbino（乌尔比诺大学纯科学与应用科学系）
