On Sept. 28, 2005, the National Cancer Institute and the National Institute of Neurological Disorders and Stroke (NINDS) created Rembrandt (Repository for Molecular BRAin Neoplasia DaTa), a joint informatics initiative to molecularly characterize a large number of primary brain tumors and to correlate those data with extensive retrospective and prospective clinical data. Understanding the biology behind these tumors and overlaying this valuable data on clinical data will provide clues to discover new therapies.

The Rembrandt brain cancer dataset includes 671 patients collected from 14 contributing institutions from 2004–2006. It is accessible for conducting clinical translational research using the open access Georgetown Database of Cancer (G-DOC) platform. In addition, the raw and processed genomics and transcriptomics data have also been made available via the public NCBI GEO repository as a super series GSE108476. Such combined datasets would provide researchers with a unique opportunity to conduct integrative analysis of gene expression and copy number changes in patients alongside clinical outcomes (overall survival) using this large brain cancer study.

The Cancer Genome Atlas (TCGA) is a comprehensive collection of multiple omics data from 33 different cancers. TCGA has two brain cancer dataset collections. One is a collection of 617 cases with grade IV gliomas referred to as TCGA-GBM. In 2015, TCGA included a cohort of lower grade glioma cases (TCGA-LGG) (https://portal.gdc.cancer.gov/projects/TCGA-LGG)¹³ that included 517 grades II and III brain cancer cases. In contrast, the Rembrandt dataset contains clinical and molecular data on 671 cases from grade II, III, and IV gliomas. The TCGA brain cancer collection was used to determine subtype classification of tumors based on multi-omics profiling of sampless^14,15.

The REMBRANDT collection is a large single study collection of brain cancers that was developed independent of TCGA efforts and provides a unique independent validation dataset for comparative analysis with TCGA. For instance Cooper et al corroborated the subtype classification obtained from the TCGA data using the Rembrandt dataset¹⁶ (https://cancergenome.nih.gov/researchhighlights/researchbriefs/corroboratesubtypes). This dataset lends itself to development of additional machine learning approaches including deep learning methods for assessing clinical relevance of biomarkers for diagnostic or therapeutic development.