Genetic predictors of 16,000 multi-omic traits and associations with breast cancer survival outcomes in the Pathways Study.

Abstract

10545 Background: Polygenic scores (PGS) enable the computation of a genetic predictor for any trait where a well-developed algorithm is available. This is appealing when such trait is not directly measured in the study population for association testing. Based on a large prospective breast cancer cohort, we calculated and investigated the prognostic value of >16,000 PGS of multi-omic traits, including plasma proteomics, plasma metabolomics, and whole-blood transcriptomics. Methods: The Pathways Study is a prospective cohort study of women with breast cancer who were enrolled soon after diagnosis in 2006-2013 at Kaiser Permanente Northern California, with ongoing follow-up. Using genome-wide genotypes from 3,995 study participants, we calculated 16,020 multi-omic PGS from the INTERVAL study ( https://www.omicspred.org/ ). We analyzed three outcomes: overall survival (OS), breast cancer specific survival (BCSS), and disease-free survival (DFS), with a median (range) follow-up time of 10.5 (0.2-14.2) years. We derived hazard ratios (HRs) and 95% confidence intervals (CIs) for one standard deviation (sd) increment in PGS from multivariable hazards models in the total study population, among those self-reported as non-Hispanic White (NHW), and by tumor estrogen receptor (ER) status in NHW women, with multiple testing corrected by a false discovery rate (FDR) of q<0.10. Results: The median age at diagnosis was 60 (23.6-94.8) years, with the majority of the study population (68%) self-identifying as NHW. Most women had ER+ tumors (83.4%), diagnosis at stages I and II (89%), and 13.3% had HER2+ tumors. The majority (60%) had lumpectomy and adjuvant therapy (44.3% radiotherapy, 47.0% chemotherapy, 74.6% hormonal therapy). Four PGS-survival outcome tests reached statistical significance with an FDR q<0.10, notably all between gene expression PGS and OS, including OPGS013029 for TINCR in NHW patients, and OPGS011920 for TTLL7 , OPGS009036 for RNF4 , and OPGS011365 for GGT7 in NHW ER- patients (Table). None of the PGS for proteomic or metabolomic traits reached this level of significance after controlling for multi-testing. Conclusions: By leveraging a large catalog of PGS of multi-omic molecular traits, we identified PGS for whole-blood RNA expression for four genes in significant associations with overall survival. Pending on validation in future studies, these genes may have prognostic value for breast cancer. PGS of multi-omics traits and associations with OS in the pathways study. Study population(race, ER status) PGS Outcome Multi-omic trait HR (95%CI) per sd increment of PGS p FDR q NHW OPGS013029 OS RNA_TINCR 0.82 (0.76, 0.89) 3.01E-06 0.05 NHW ER- OPGS011920 OS RNA_TTLL7 1.40 (1.22, 1.60) 8.45E-07 0.01 NHW ER- OPGS009036 OS RNA_RNF4 1.54 (1.28, 1.85) 4.07E-06 0.03 NHW ER- OPGS011365 OS RNA_GGT7 1.54 (1.27, 1.88) 1.60E-05 0.09