Abstract 2406: A general probabilistic algorithm to predict de novo mutations in familial diseases as demonstrated in Li-Fraumeni Syndrome

Abstract

Abstract Purpose: Germline variants (mutations) occurring in the carrier but not in the carrier's parents are defined as de novo mutations (DNMs). DNMs have been increasingly recognized as causal factors for rare diseases. Given this important role, the ability to identify DNM carriers, scarcely distributed among all mutation carriers, will allow researchers to study the likely distinct molecular mechanisms of DNMs in these diseases. However, testing the mutation status of the parents is often impractical in retrospective studies due to lack of blood samples. Methods: To fill this need, using Li-Fraumeni syndrome (LFS) as a representative of a rare inherited syndrome, we developed a method called Famdenovo to predict the de novo status (de novo or familial) of germline mutations, such as in TP53. We have collected a total 324 LFS family pedigrees with confirmed TP53 germline mutations from four medical centers in the US, with 186 pedigrees having TP53 genetic testing results available in at least one full trio in the pedigree, serving as a validation set. We further apply Famdenovo to the remaining 138 families (discovery set) to identify individuals who may carry a de novo mutation in TP53. Results: In the validation set, the area under the ROC curve (AUC) of Famdenovo is 0.95 (95% CI: [0.92, 0.98]), suggesting excellent ability for discrimination, and the ratio of the observed to expected (OE) is 1.32 suggesting good concordance. In the discovery set, we predict that an additional 40 individuals (95% CI: [30, 50]) are DNM carriers, which increases the total number of DNM carriers to 82. The corresponding predicted DNM rate is 28.9%, also consistent with what is observed in the validation set. Across the validation and discovery sets, we observed similar distributions of ages-of-onset for DNM carriers at specific cancer sites: breast, brain, leukemia, osteosarcoma, soft tissue sarcoma, lung, and adrenal cortical carcinoma. Interestingly, Lung cancer only occurred in female DNM carriers. Among TP53 hotspot mutations, R248Q is most enriched in DNMs. Conclusions: Our new statistical method Famdenovo provides the probability of a person carrying a de novo germline mutation in rare inherited syndromes when the mutation status of both parents is not available. Famdenovo is freely available as an R package from http://bioinformatics.mdanderson.org/main/Famdenovo. The LFS family cohorts not only served as a validation set for the accuracy of Famdenovo in predicting DNMs, but also benefited from using Famdenovo to discover additional individuals carrying DNMs in TP53. The computer-based identification of DNM carriers in TP53, who are otherwise hidden in a wide population, enabled epidemiological interpretation of the cancer outcomes for DNM carriers. Famdenovo is a general tool and can be applied to other cancer genes where there is a good understanding of the penetrance of the associated disease phenotype. Citation Format: Xuedong Pan, Fan Gao, Elissa B. Dodd, Jasmina Bojadzieva, Phuong L. Mai, Valen E. Johnson, Kristin Zelley, Kim E. Nichols, Judy E. Garber, Sharon A. Savage, Louise C. Strong, Wenyi Wang. A general probabilistic algorithm to predict de novo mutations in familial diseases as demonstrated in Li-Fraumeni Syndrome [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 2406.