Abstract A few germline pathogenic structural variants (SV) have been identified in cancer predisposition syndromes, e.g., MSH2 inversion in Lynch syndrome. The advent of short-read whole-genome sequencing (WGS) has facilitated the detection of SVs. However, a pitfall of this sequencing methodology is the inability to capture all SVs, given the reads do not map well to low complexity regions, and current algorithms used to identify SVs from short-read data have very low sensitivity and very high false-positive rates. One of the objectives of this fellowship is to optimize and implement an SV calling pipeline that utilizes multiple algorithms to increase the sensitivity and specificity of variant identification from germline short-read WGS trios. This pipeline is currently being developed and tested on pediatric cancer patients enrolled in the NIH-funded Baylor Advancing Sequencing into Childhood Cancer Care (BASIC3) exome study, which consists of ethnically and racially diverse pediatric patients with solid (CNS and non-CNS) tumors. This BASIC3 subset includes 63 proband/parent trios who have subsequently undergone germline short-read WGS. Five SV callers are run and then filtered using either a percent reciprocal overlap filter or a proposed Artificial Intelligence-based proximity graph filter to identify de novo and inherited SVs. This cohort has recently been selected for a long-read sequencing pilot, data which serves as the gold standard for SV detection. Comparison of short-read WGS generated SV calls with the long- read data will help determine the sensitivity and specificity of variant calls and further improve the short read pipeline (Aim 1A). Once optimized this short-read WGS SV pipeline will be applied to the Kids First Genetics of Embryonal and Alveolar Rhabdomyosarcoma cohort (n=900), as an independent assessment of the method and also to identify recurrent germline SVs in this cancer which has not yet been well characterized (Aim 1B). The second training and scientific objective is to explore the functional effect of a novel de novo SV identified through the initial analysis. A de novo germline duplication of the Prostaglandin Reductase 2 (PTGR2) enhancer and promoter region was found in a pediatric Posterior Fossa subtype A (PF–A) ependymoma patient. Transcriptome data from the patient’s tumor revealed increased PTGR2 expression. The PTGR2 protein converts 15-keto- prostaglandin E2 to 15-keto-13,14-dihydro-PGE2. Studies in adult malignancies suggest that increased 15-keto- 13,14-dihydro-PGE2 is associated with cancer risk potentially through the STAT3 signaling pathway. Increased STAT3 signaling is reported as a distinct feature of PF-A ependymoma. The identified partial duplication of PTGR2 will be engineered into ependymoma progenitor cells, to functionally assess the effects on STAT3 signaling, cell proliferation, and DNA synthesis. In parallel, further analysis of PTGR2 expression and structural variants will be analyzed in pe...