Abstract

Abstract Homologous recombination repair deficiency (HRD) and mutations in genes involved with homologous recombination repair (HRR) are associated with sensitivity to platinum-based chemotherapy, topoisomerase 1 inhibitors and inhibitors targeting DNA repair pathway. Use of the Myriad HRD test for HRD determination in preclinical models is not feasible globally. To address this challenge, we developed an approach to generate HRD scores from whole exome sequencing (WES) data and applied it to assess HRD status in patient-derived xenograft (PDX) models. We used python package CNVkit for segmentation and gene-level/allele-specific copy number calling with customized settings according to CNVkit documentation. The reference was constructed using the WES BED file and hg38 reference genome when there are no paired control samples available. Next R package scarHRD was used to determine the HRD score that is the sum of loss of heterozygosity (LOH), telomere allele imbalances (TAI), and large fragment migrations (LST). The method was applied to the WES of 25 Xentech triple negative breast cancer (TNBC) PDX models to derive HRD scores. The samples were also separately profiled using Myriad’s MyChoice CDx assay. The two scores showed moderate correlation with Pearson r = 0.56. However, when n = 6 PDX models with low coverages (average depth <100x) were excluded, the Pearson r improved to r = 0.76. We also found that the absolute value of HRD score calculated by CNVkit/scarHRD is lower than that of Myriad HRD score. Furthermore, we evaluated the genomic landscape in this cohort of TNBC PDX models for the selected genes of HRR pathway (BRCA1, BRCA2, ATM, BRIP1, PALB2, RAD51C, BARD1, CDK12, CHEK1/2, FANCL, RAD51B, RAD51D, RAD54L). Sixteen models found to carry HRR mutations, with ATM and BRCA1 mutations being the most common ones, being found in 9 and 5 of 25 models, respectively. Although HRRm positive (HRRm+) models have higher HRD score for both Myriad HRD score and HRD score generated with CNVkit/scarHRD approach, the association between HRR (HRRm+/HRRm-) and Myriad HRD (HRD+/HRD- defined by HRD threshold of 42) is not statistically significant (Fisher’s exact test p value = 0.1998, odds ratio=3.5), likely due to small cohort sizes. These results suggest that combination of CNVkit and scarHRD for HRD score calculation may be an alternative approach for HRD identification in preclinical PDX models. Citation Format: Jixin Wang, Krista Kinneer, Wenyan Zhong, Zhongwu Lai. The development of homologous recombination repair deficiency (HRD) score methodology using WES data and its application in patient-derived xenograft models [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 6553.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call