Abstract 2464: Gabriella Miller Kids First Data Resource Center: Harmonizing clinical and genomic data to support childhood cancer and structural birth defect research

Abstract

Childhood cancers and structural birth defects share a common context of altered developmental biology, but the potential role of shared, genetic alterations and/or pathways across pediatric cancers and birth defects is not well explored. It is increasingly critical that genomic data are paired with high-quality clinical data to drive translational research by elucidating the relationship between genomic alterations, treatments, outcomes, and other phenotypic characteristics. The NIH Common Fund Gabriella Miller Kids First Program represents a first-in-kind national, collaborative initiative focused on large-scale clinical and genomic data sharing for childhood cancers and structural birth defects. As part of this program, the Kids First Data Resource Center (DRC) is charged with empowering collaborative discovery across Kids First datasets. Through newly developed cloud-based platforms, researchers will be able to rapidly and interactively access standardized and harmonized clinical and genomic data. A better understanding of common developmental programs could spur advancements in prevention, detection, and therapeutics that will improve the outcomes of affected children and families. Approximately 8,000 patient samples were available at the launch of the Kids First DRC portal, with an initial focus on whole genome sequencing (WGS) of trios and families. More than 25,000 WGS are expected to be processed by 2019, making the DRC one of the largest pediatric data resources of its kind across a diversity of diseases. The rise of cloud-based computing has greatly reduced the burden on the researcher of large-scale genomic harmonization. In normal operations, the DRC is capable of running two hundred workflows simultaneously with considerable scalability on demand. Additionally, there is a strong focus on harmonizing and structuring seemingly disparate clinical and phenotypic data types to make them more interoperable, discoverable and reusable by using ontologies. The data in the DRC is expertly curated and mapped to existing data standards, including NCI Thesaurus (NCIt), Human Phenotype Ontology (HPO), Monarch Disease Ontology (MONDO), and Uber-anatomy Ontology (Uberon). This allows for increased interoperability and semantic structure of the data. For example, MONDO integrates numerous disease terminologies into a single merged ontology, including the NCIt. The combination of harmonized genomic and clinical data across pediatric cancers and structural birth defect provides a key foundation for exploring and developing new methods to better understand the relationships between germline variants, cancer risk, and associated treatments and outcomes. Community standardization of this modeling is ongoing as part of GA4GH, and is critical for implementation of improved interpretation in EHR systems, for example via HL7 FHIR. Citation Format: Allison P. Heath, Deanne M. Taylor, Yuankun Zhu, Pichai Raman, Jena Lilly, Phillip Storm, Angela J. Waanders, Vincent Ferretti, Christina Yung, Michele Mattioni, Brandi Davis-Dusenbery, Zachary L. Flamig, Robert Grossman, Samuel L. Volchenboum, Sabine Mueller, Javad Nazarian, Nicole Vasilevsky, Melissa A. Haendel, Adam Resnick. Gabriella Miller Kids First Data Resource Center: Harmonizing clinical and genomic data to support childhood cancer and structural birth defect research [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2464.

Towards Self-Describing and FAIR Bulk Formats for Biomedical Data

“…We also performed several experiments using the structured project data from three large scale cloud-based data platforms: the KidsFirst Data Resource [11], the NHLBI BioData Catalyst system [14], and the NCI Genomic Data Commons (GDC) [10]. We took: 1) PostgreSQL dumps and 2) PFB exports from each system.…”

trioPhaser: using Mendelian inheritance logic to improve genomic phasing of trios

“…In addition to validating trioPhaser with GIAB data, we ran trioPhaser on 50 trios where the child in each trio had been diagnosed with neuroblastoma. These trios and their associated gVCF files are available through the Gabriella Miller Kids First Data Resource Center [ 17 ]. The gVCF files were generated using WGS data.…”

Germline predisposition to pediatric Ewing sarcoma is uniquely characterized by inherited pathogenic variants in DNA damage repair genes

“…Sequenced samples from the discovery cohort came from four data sources: St. Jude Cloud (Pediatric Cancer Genome Project, St. Jude Lifetime, Genomes for Kids, and Childhood Cancer Survivor Study; n = 1033) 42–46 , dbGaP study “Genomic Sequencing of Ewing Sarcoma” (phs000804.v1.p1; n = 26) 47,48 , dbGaP study “Osteosarcoma Genomics” (phs000699.v1.p1, n = 58) 49 , and ICGC study “Bone Cancer – UK” (BOCA-UK, n = 63) 50 . Sequenced samples from the Ewing sarcoma validation and trio cohorts were from the Gabriella Miller Kids First “Ewing Sarcoma – Genetic Risk” study (phs001228.v1.p1) 51 . The control cohorts came from the following sources: Autism Sequencing Consortium (dbGaP phs000298), Framingham Cohort (dbGaP phs000007), Multi-Ethnic Study of Atherosclerosis (dbGaP phs000209), Lung Cohort (dbGaP phs000291), in-house collection of exomes from National Heart, Lung, and Blood Institute “Grand Opportunity” Exome Sequencing Project (NHLBI GO-ESP), and the 1000 Genomes Project 52 .…”