Abstract 1655: A structural approach for prioritizing variants in personalized Oncology

Abstract

Abstract In personalized oncology, treatment strategies take individualized genomic aberrations into consideration and relies heavily on how the cancer genomes are interpreted. With immense efforts put into cohort studies, knowledge of pathogenicity at the gene level had been accumulated. On the other hand, the increasing amount of personalized clinical trials and the accompanying need of interpreting pathogenecity at the sub-gene level emphasize the importance of resolving heterogeneous significance within genes. There had been some initiatives where people identified mutation hotspots within genes, looking at either highly recurrent single residues or mutation clusters. The hotspot clusters can be inferred from protein sequences as well as from protein structures, where residues distant in sequence may be brought together through protein folding. Nonetheless, these information are not well utilized in routine clinical settings yet. In this study, we aim to build a consensus list of prevalent mutational events for molecular stratification of patients. This is achieved by collecting sequence and structural mutation hotspots from different algorithms, and further combine structural information to identify their neighboring residues in protein structures. Specifically, around 1400 residue hotspots are collected from CancerHotspots and DominoEffect. An additional of around 250 structural hotspots are identified by three structural algorithms, namely 3DHotSpots, HotMAPS and HotSpot3D. These results are mostly derived from TCGA PanCancer cohorts, with CancerHotspots included more other studies in the analysis. With this consensus information in place, we further set up the mechanism to map from genomic variants to protein sequence and protein structures. This enables us to annotate the variants found in patient genome and see whether they co-localize with hotspot residues, in their structural neighborhood, or in structural informative domains like interaction interfaces. In summary, a higher prevalence of variants might imply higher significance and more well-studied roles than other low prevalent mutant alleles. Transferring this knowledge from the public cohorts helps researchers quickly prioritize variants found in patient genomes. Annotating structural significance of variants further provides us with more confidence in identifying driving events. With these procedures automated, the goal is to help facilitate clinical evaluations to meet the growing number of patients recruited to personalized oncology trials. Citation Format: Siao-Han Wong, Brors Benedikt. A structural approach for prioritizing variants in personalized Oncology [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 1655.