PROJECT SUMMARY Circular extrachromosomal DNA (ecDNA) is a special type of structural variant and a common vehicle of high oncogene amplification in cancers. Due to the lack of centromeres, ecDNAs can segregate unevenly to daughter cells during cell division, leading to substantial cell-to-cell variability and a high copy number of ecDNAs in some cells. ecDNAs have been found to be important prognostic markers in some cancers. Mechanistically, ecDNAs can create many copies of proto-oncogenes and up-regulate their expression by providing a high chromatin accessibility environment and sometimes rewiring of enhancer targeting. Identifying ecDNAs from sequencing data is challenging due to complex rearrangements of DNA segments in the ecDNA, which make read alignment and inference of segment connection structures non-trivial. In addition, other types of non-ecDNA genomic rearrangements can also confuse the inference process. To address these issues, we have previously developed and thoroughly tested a computational pipeline that can accurately identify ecDNA from whole-genome sequencing data. Using the pipeline, we have discovered that ecDNA is associated with higher cancer recurrence rate and poorer patient survival in medulloblastoma. This project aims at facilitating the investigations of the prevalence, prognostic value, and functional significance of ecDNA in other childhood, adolescent and young adult (AYA) cancers. We will collect sequencing data from these cancer types for more than 3,500 tumor samples. We will apply our established computational pipeline to identify ecDNAs from all these samples. We will then build a web portal such that all these results are readily accessible to the public. The web portal will also provide functionalities for performing various kinds of comparisons between groups of samples, patients, or ecDNAs and visualize the results using standard plots. For example, users can define groups of patients with different sex or age/ethnicity and compare how their cancer recurrence rates and survival time are affected by presence of ecDNA or specific genes contained in the ecDNA. Such comparisons can identify groups of patients particularly susceptible to ecDNA-associated poor prognosis and suggest ways to perform patient stratification. In our project we will also systematically perform this type of comparisons for groups defined by all possible single- and double-variable combinations, with the statistical significance of all comparison results carefully evaluated using rigorous procedures. All these results will also be made accessible on our web portal. The statistically significant results can be used by anyone studying childhood and AYA cancers to form concrete hypotheses and investigate further. Overall, our resource will substantially lower the barrier of studying ecDNAs in childhood and AYA cancers. Our resource will also contribute toward the Childhood Cancer Data Initiative by supplying information about the no...