NEW YORK – By digging into tissue-specific and cross-tissue transcriptomic features, members of two independent research teams have tracked down new and known genes with apparent ties to breast and/or ovarian cancer risk.
In a study published in the American Journal of Human Genetics on Friday, public health science, genetic medicine, and hematology and oncology researchers from the University of Chicago shared findings from a transcriptome-wide association study (TWAS) that took splicing-TWAS data from 11 tissues into account when searching for breast cancer susceptibility genes.
"[W]e performed a joint splicing-TWAS that combined information from multiple excised introns in each gene across multiple tissues that are potentially relevant to breast cancer," co-senior and co-corresponding authors Hae Kyung Im and Dezheng Huo and their colleagues wrote in their paper.
The team uncovered 240 genes with significant ties to breast cancer with a multi-tissue joint splicing TWAS that combined splicing prediction profiles from version 8 of the Genotype-Tissue Expression (GTEx) project with genome-wide association study summary statistic data from a meta-analysis that included almost 424,700 European ancestry women with or without the disease who were enrolled through the Breast Cancer Association Consortium (BCAC) or the UK Biobank project.
Nine more associated genes turned up with a breast tissue-centered analysis, the researchers explained, bringing the total to 249 genes. That set encompassed genes such as TRIM4, GJC3, AZGP1, AFF1, SRP54, and ZNF that fell at sites near breast cancer-associated variants uncovered using a GWAS approach.
They noted that 88 of the breast cancer-associated genes appeared to be novel, while 110 genes they uncovered had not been picked up in prior TWAS analyses that used the same GWAS summary statistic and GTEx data sources but focused on gene expression profiles.
Based on these and other results, the authors argued that "for many genes, expression quantitative trait loci (eQTL) did not show a significant impact on breast cancer risk, whereas splicing quantitative trait loci (sQTL) showed a strong impact through intron excision events."
"These findings suggested that sQTL-based, splicing-based TWASs might provide different information from eQTL-based, expression-based TWASs for breast cancer risk," they added, "and might reveal new insights into the genetic etiology of breast cancer."
In a related study appearing in the American Journal of Human Genetics on Wednesday, a team from Emory University, St. Jude Children's Research Hospital, Cedars-Sinai Medical Hospital, and the University of Cambridge used a TWAS approach to search for breast or ovarian cancer-related genes, focusing on cis- and trans-acting regulatory variants influencing the expression of genes found near or far from the eQTLs, respectively.
After training their imputation models with whole-genome sequencing and RNA-seq data on breast and ovarian tissues analyzed for GTEx, the investigators went on to search for genes associated with breast or ovarian cancers or cancer subtypes using summary statistic data from BCAC or Ovarian Cancer Association Consortium GWAS efforts.
"This approach both improves prediction of [genetically regulated expression] as well as enhances detection of genes that influence phenotype through distal transcriptional regulation," senior and corresponding author Michael Epstein and co-authors wrote.
After highlighting 101 genes with apparent ties to breast cancer and eight ovarian cancer-associated genes, the team went on to validate a subset of the findings with summary statistics from another GWAS, analyzed in combination with breast tissue expression and genotyping data from the Cancer Genome Atlas project.
"Our findings replicate several established cancer risk loci and suggest several candidate trans-eQTL-driven genes not discovered by a standard TWAS approach that models cis-SNPs only," the authors reported, adding that the study "provides insight into the eQTL genetic architecture of breast and ovarian tissue and leverages trans-genome regulation of expression in these tissue for improved TWASs of breast and ovarian cancer."