Abstract for presentation at 11th International Congress of Human Genetics

A genome-wide association study of early-onset BRCA1/2-negative invasive breast cancer among non-Hispanic Caucasian women

  • Muhammad Kibriya, Mailman School of Public Health, Columbia University, New York, USA., United States
  • Maria Argos, Mailman School of Public Health, Columbia University, New York, USA., United States
  • Farzana Jasmine, Mailman School of Public Health, Columbia University, New York, USA., United States
  • Irene Andrulis, University of Toronto, ON, Canada, Canada
  • Esther John, Northern California Cancer Center, CA, United States
  • Jenny Chang-Claude, German Cancer Research Center, Heidelberg, Germany
  • Habibul Ahsan, Mailman School of Public Health, Columbia University, New York, USA., United States

    • Genome-wide association (GWA) analysis is now possible with the availability of high-density oligonucleotide microarray platform containing a large number of single nucleotide polymorphisms (SNP). We used the early access Affymetrix Mendel Nsp 250K chips in a GWA case-control study to identify SNPs associated with breast cancer. We included 30 non-Hispanic Caucasian BRCA1/2-negative population-based incident cases of invasive breast cancer aged <40 years from 3 centers (USA, Canada and Germany) and 30 population controls individually matched on age, ethnicity and center. The overall call rate was 89.04% ± 3.51% for controls and 9.12% ± 4.16% for cases without any significant difference between the centers. After checking for HWE and genotyping errors, SNP by SNP comparison was made between cases and controls for all the 224940 genotyped SNPs using conditional logistic regression (CLR) models with adjustment for the matching factors. For initial screening purposes, SNPs showing a p-value =<0.001 based on a 1-df trend test from the CLR models were selected for further in-depth exploration. For each selected SNP, the cytoband region was mapped and signals were picked with Sime’s method using a window of 3-5 with adjustment for multiple comparisons using both FDR and Bonferroni methods. Then association with disease was checked using haplotype of 2 consecutive SNPs (HAP2), 3 consecutive SNPs (HAP3) and so on in the same region. We also identified the tag SNPs by Carlson method and traced the SNPs of interest within / around the tag SNPs. If any of the SNPs within the HAP2/HAP3 was proxy for other tag-SNP, then that was dropped and new haplotype and diplotype were constructed taking the next nearby available tag SNP and looked for association with disease. Finally, we checked if the associations were significant in multivariate CLR models adjusted for matching factors and established breast cancer risk factors. Using the above approach, we have identified a number of potentially breast cancer associated loci including some novel loci and some previously reported loci. Details findings of these promising breast cancer loci identified in this GWA study will be presented in the meeting.

    Conference Organiser - ICMS Pty Ltd