SimBioSys PhenoScope: A Cloud-Enabled, Web-Based Portal for Raw and Processed Cancer Phenotype Data Spanning Scales from Genomes to Tissues

虽然基因组学方法产生了大量关于癌症潜在原因和种类的信息，但基因只是影响肿瘤表型的一个因素。癌症之间的异质性和个体肿瘤内部的差异继续挑战着开发有效治疗方法的努力。正在产生大量数据，如单细胞测序、数字病理学和医学成像，这些数据捕获了多个尺度上的异质性。然而，数据是孤立的，规模之间的联系仍然难以捉摸。我们建议开发SimBioSys PhenoScope，这是一个新的研究工具，可以协调这些不同的数据类型，并提供基因组、细胞/途径、微观组织环境和组织尺度之间的联系。PhenoScope结合了最先进的机器学习、降维技术、新的时空模拟算法以及对公共数据库的支持，将提供一种评估癌症表型行为因素的新手段。作为一个探索性数据平台，该工具将依靠新的尺度间分析提供新的多尺度可视化，学术和制药研究人员都可以使用这些分析来生成新的药物靶点、给药方案和研究目标的假设。这项技术使研究人员和临床医生能够以新的视角研究癌症，并引领着即将到来的表型组学时代。

While genomic methods yield a plethora of information on the underlying cause and variety of cancers, genotype is only one factor contributing to the observed phenotype of a tumor. Heterogeneity among cancers and differences within individual tumors continue to challenge the efforts to develop effective therapies. A multitude of data such as single cell sequencing, digital pathology, and medical imaging are being generated that captures heterogeneity across multiple scales. However, data is siloed and connections between scales remains elusive. We propose the development of SimBioSys PhenoScope, a novel research tool to harmonize these disparate data types and provide connections between the genomic, cellular/pathway, microscopic tissue environment, and tissue scales. Combining state-of-the art machine learning, dimensionality reduction techniques, novel spatio-temporal simulation algorithms, and support for public data repositories, PhenoScope will provide a new means of assessing factors contributing to a cancer's phenotypical behavior. As an exploratory data platform, the tool will provide novel multi-scale visualizations by relying on the novel analyses between scales, that could be used by academic and pharmaceutical researchers alike to generate hypotheses for new drug targets, dosing regimens, and research targets. This technology enables researchers and clinicians to study cancer in a new light, and lead the way in to the upcoming phenomics era.