Integrated computational modeling of multi-scale biomedical data

胰腺癌是一种致命的疾病，由于诊断的晚期和缺乏有效的治疗方法。拟议的研究计划将开发新的方法来分析和整合多尺度的生物医学数据，以更好地阐明胰腺癌的潜在机制，其中单独的突变分析无法定义完整的生物景观。该计划旨在通过整合诊断成像，组织病理学切片，基因组数据和其他来源的基于组学的分析来弥合这一关键障碍，以提供一种统一的，令人信服的胰腺癌表征方法。 机器学习算法的实质性进步，结合诊断成像分辨率的提高，显微镜载玻片的常规数字化以及下一代测序的涌入，创造了该计划所利用的无与伦比的机会。诊断成像，分子分析和组织学成像比任何单独的方式创建一个更完整的生物图片。我们的长期目标是利用数据来改变患者的治疗方式，利用新的计算方法来实现精准医疗的承诺-这是现代癌症护理的最大挑战。该计划对高素质人员的培训将包括机器学习，生物医学数据科学和计算建模方面的经验。五名博士，三名硕士和四名本科生将在拟议的计划中接受培训。在机器学习和生物医学数据科学方面接受过培训的HQP在加拿大需求量很大，学员将在加速最先进的方法方面发挥关键作用，这些方法将通过提供个性化治疗从根本上改变癌症患者的护理，降低加拿大医疗保健的成本。

Pancreas cancer is a uniformly deadly disease due to the late stage at diagnosis and the lack of effective therapies. The proposed research program will develop new methods to analyze and integrate biomedical data at multiple scales to better elucidate the underlying mechanisms of pancreatic cancer where mutational analysis alone fails to define complete biological landscapes. The proposed program aims to bridge this critical barrier by integrating omics-based analyses of diagnostic imaging, histopathology slides, genomic data, and other sources to provide a unified, cogent approach to pancreatic cancer characterization. Substantial advances in machine learning algorithms combined with improvements in the resolution of diagnostic imaging, routine digitization of microscopy slides, and the influx of next generation sequencing has created an unrivaled opportunity that this program exploits. Diagnostic imaging, molecular profiling, and histologic imaging create a more complete biological picture than any individual modality. Our long-term goal is to harness data to change the way that patients are treated, utilizing novel computational methods to realize the promise of precision medicine - the greatest challenge in modern cancer care. Training of highly qualified personnel in this program will include experience in machine learning, biomedical data science, and computational modeling. Five PhD, three Master's, and four undergraduate students will receive training in the proposed program. HQP trained in machine learning and biomedical data science are in high demand in Canada and trainees will play a critical role in accelerating state-of-the-art approaches that will radically transform care for cancer patients by informing individualized treatments, at a reduced cost to Canadian healthcare.