The Glioma Molecular Diagnostic Initiative: Characterizing Brain Tumor Data

美国国家癌症研究所（NCI）神经肿瘤学分支（NOB）在过去六年中收集和描述了近1000个不同级别和组织病理学分类的胶质瘤（少突胶质细胞瘤、星形细胞瘤、室管膜瘤），这是一项由H. Fine构思、撰写并主持的多机构研究（NABTC协议01-07）。这项研究被广泛地称为神经胶质瘤分子诊断计划（GMDI），旨在获得大量新采集的肿瘤样本的DNA和RNA分子数据，这些样本以标准化的方式收集、处理和分析，以便进行大规模的交叉样本分析。此外，样品收集伴随着仔细和前瞻性的临床数据采集，允许前所未有的丰富匹配的分子和临床数据允许各种各样的分析。与此同时，NOB和NCI生物信息学中心也进行了类似的广泛合作，为GMDI数据分析创建了一个数据仓库和用户友好的跨数据平台生物信息学工具棚，称为REMBRANDT。由此产生的公共数据库几乎是前所未有的，不仅在肿瘤学领域，而且在所有分子医学领域都是如此，它证明了校内项目的可能性，以及它如何影响整个国家癌症研究事业。GMDI在回顾性阶段共收集了268个新鲜冷冻肿瘤（全部来自亨利福特医院，无生殖系DNA），在前瞻性阶段收集了670个新鲜冷冻肿瘤（来自各种机构），总共938个肿瘤代表了所有胶质瘤类型和级别。其中，756份（80.5%）已被处理并杂交到Genechip Human Genome U133 Plus 2.0表达阵列，而729份（77.7%）已被杂交到Affymetrix Genechip Human Mapping 100K阵列进行基因组鉴定。意识到大多数来找我们的患者没有新鲜的冷冻组织，我们花了一年多的时间开发和优化一种方法，使我们能够从FFPE中提取高质量的DNA，从而对这些患者进行基因组分析。到目前为止，已经收集了165例这样的病例，其中117例（71%）已经使用Affymetrix基因芯片人类图谱250K StyI阵列和我们的FFPE DNA恢复方案进行了基因组表征。细胞系：除了对GMDI患者的样本进行表征外，微阵列组还对许多人类和犬胶质瘤起始细胞/胶质瘤干细胞（GIC/GSC）系以及实验室生产的许多犬和鼠正常神经干细胞（NSC）系进行了基因组规模分析。这种表征既在原代细胞水平上（包括通过传代的进化），也评估不同处理（分化、动物传代、药物处理等）对细胞生物学行为的影响。总共完成了202个表达序列（来自人类、小鼠和犬）以及40个人类SNP序列。美国国家癌症研究所（NCI）神经肿瘤学分支（NOB）在过去六年中收集和描述了近1000个不同级别和组织病理学分类的胶质瘤（少突胶质细胞瘤、星形细胞瘤、室管膜瘤），这是一项由H. Fine构思、撰写并主持的多机构研究（NABTC协议01-07）。这项研究被广泛地称为神经胶质瘤分子诊断计划（GMDI），旨在获得大量新采集的肿瘤样本的DNA和RNA分子数据，这些样本以标准化的方式收集、处理和分析，以便进行大规模的交叉样本分析。此外，样品收集伴随着仔细和前瞻性的临床数据采集，允许前所未有的丰富匹配的分子和临床数据允许各种各样的分析。与此同时，NOB和NCI生物信息学中心也进行了类似的广泛合作，为GMDI数据分析创建了一个数据仓库和用户友好的跨数据平台生物信息学工具棚，称为REMBRANDT。由此产生的公共数据库几乎是前所未有的，不仅在肿瘤学领域，而且在所有分子医学领域都是如此，它证明了校内项目的可能性，以及它如何影响整个国家癌症研究事业。GMDI在回顾性阶段共收集了268个新鲜冷冻肿瘤（全部来自亨利福特医院，无生殖系DNA），在前瞻性阶段收集了670个新鲜冷冻肿瘤（来自各种机构），总共938个肿瘤代表了所有胶质瘤类型和级别。其中，756份（80.5%）已被处理并杂交到Genechip Human Genome U133 Plus 2.0表达阵列，而729份（77.7%）已被杂交到Affymetrix Genechip Human Mapping 100K阵列进行基因组鉴定。意识到大多数来找我们的患者没有新鲜的冷冻组织，我们花了一年多的时间开发和优化一种方法，使我们能够从FFPE中提取高质量的DNA，从而对这些患者进行基因组分析。到目前为止，已经收集了165例这样的病例，其中117例（71%）已经使用Affymetrix基因芯片人类图谱250K StyI阵列和我们的FFPE DNA恢复方案进行了基因组表征。细胞系：除了对GMDI患者的样本进行表征外，微阵列组还对许多人类和犬胶质瘤起始细胞/胶质瘤干细胞（GIC/GSC）系以及实验室生产的许多犬和鼠正常神经干细胞（NSC）系进行了基因组规模分析。这种表征既在原代细胞水平上（包括通过传代的进化），也评估不同处理（分化、动物传代、药物处理等）对细胞生物学行为的影响。总共完成了202个表达序列（来自人类、小鼠和犬）以及40个人类SNP序列。

The Neuro-Oncology Branch (NOB) of the National Cancer Institute (NCI) for the past six years has been collecting and characterizing nearly one thousand gliomas of every grade and histopathological class (oligodendrogliomas, astrocytomas, ependymomas) under a multi-institutional study conceived of, written by and chaired by H. Fine (NABTC protocol 01-07). This study is broadly termed the Glioma Molecular Diagnostic Initiative (GMDI), and was designed to obtain a large amount of molecular data on DNA and RNA of freshly collected tumor samples that were collected, processed and analyzed in a standardized fashion to allow for large-scale cross sample analysis. Moreover, the sample collection is accompanied by careful and prospective clinical data acquisition, allowing an unprecedented wealth of matched molecular and clinical data permitting a wide variety of analyses. The effort has been accompanied by a similarly extensive collaborative effort between the NOB and the NCI Center for Bioinformatics to create a data warehouse and user-friendly cross-data platform bioinformatics tool shed for GMDI data analyses known as REMBRANDT. The resultant public database is nearly unprecedented, not only in oncology, but in all of molecular medicine and is a testament to what is possible within the intramural program and how it can impact on the overall national cancer research enterprise.GMDI has accrued a total of 268 fresh frozen tumors in the retrospective phase (all from the Henry Ford Hospital, without germline DNA) and 670 fresh frozen tumors in the prospective phase (from a variety of institutions) for a total of 938 tumors representing all glioma types and grades. Of these, 756 (80.5%) have been processed and hybridized to the Genechip Human Genome U133 Plus 2.0 Expression Arrays, while 729 (77.7%) have been hybridized to the Affymetrix Genechip Human Mapping 100K arrays for genomic characterization. Realizing that a majority of patients who come to us do not have obtainable fresh frozen tissue we spent over a year developing and optimizing a methodology that allows us to extract high quality DNA from FFPE enabling genomic analyses of these patients. To date, one hundred and sixty five such cases have been collected, and of these, 117 (71%) have been genomically characterized using Affymetrix Genechip Human Mapping 250K StyI arrays and our FFPE DNA restoration protocol. *Cell Lines: In addition to characterizing the samples from patients enrolled in GMDI, the microarray group has generated genomic-scale analyses of the many human and canine glioma initiating cells/glioma stem cells (GIC/GSC) lines, as well as many canine and murine normal neural stem cell (NSC) lines produced in laboratory. This characterization is both at the primary cell level (including evolution through passages) as well as evaluation of the impact of different treatments (differentiation, animal passages, drug treatment, etc) on the biological behavior of the cells. In all, 202 expression arrays have been completed (from human, mouse and canine sources) as well as 40 human SNP arrays.The Neuro-Oncology Branch (NOB) of the National Cancer Institute (NCI) for the past six years has been collecting and characterizing nearly one thousand gliomas of every grade and histopathological class (oligodendrogliomas, astrocytomas, ependymomas) under a multi-institutional study conceived of, written by and chaired by H. Fine (NABTC protocol 01-07). This study is broadly termed the Glioma Molecular Diagnostic Initiative (GMDI), and was designed to obtain a large amount of molecular data on DNA and RNA of freshly collected tumor samples that were collected, processed and analyzed in a standardized fashion to allow for large-scale cross sample analysis. Moreover, the sample collection is accompanied by careful and prospective clinical data acquisition, allowing an unprecedented wealth of matched molecular and clinical data permitting a wide variety of analyses. The effort has been accompanied by a similarly extensive collaborative effort between the NOB and the NCI Center for Bioinformatics to create a data warehouse and user-friendly cross-data platform bioinformatics tool shed for GMDI data analyses known as REMBRANDT. The resultant public database is nearly unprecedented, not only in oncology, but in all of molecular medicine and is a testament to what is possible within the intramural program and how it can impact on the overall national cancer research enterprise.GMDI has accrued a total of 268 fresh frozen tumors in the retrospective phase (all from the Henry Ford Hospital, without germline DNA) and 670 fresh frozen tumors in the prospective phase (from a variety of institutions) for a total of 938 tumors representing all glioma types and grades. Of these, 756 (80.5%) have been processed and hybridized to the Genechip Human Genome U133 Plus 2.0 Expression Arrays, while 729 (77.7%) have been hybridized to the Affymetrix Genechip Human Mapping 100K arrays for genomic characterization. Realizing that a majority of patients who come to us do not have obtainable fresh frozen tissue we spent over a year developing and optimizing a methodology that allows us to extract high quality DNA from FFPE enabling genomic analyses of these patients. To date, one hundred and sixty five such cases have been collected, and of these, 117 (71%) have been genomically characterized using Affymetrix Genechip Human Mapping 250K StyI arrays and our FFPE DNA restoration protocol. *Cell Lines: In addition to characterizing the samples from patients enrolled in GMDI, the microarray group has generated genomic-scale analyses of the many human and canine glioma initiating cells/glioma stem cells (GIC/GSC) lines, as well as many canine and murine normal neural stem cell (NSC) lines produced in laboratory. This characterization is both at the primary cell level (including evolution through passages) as well as evaluation of the impact of different treatments (differentiation, animal passages, drug treatment, etc) on the biological behavior of the cells. In all, 202 expression arrays have been completed (from human, mouse and canine sources) as well as 40 human SNP arrays.