Building genomic datasets (i.e. genomewide association studies, GWAS) for childhood cancer is a lot of work. Childhood cancer is thankfully rare, so often GWAS datasets must be cobbled together with samples from many sources. Once these precious datasets are assembled it makes sense to “squeeze the juice” out of them through integrative analysis.
In genomics research, integrative studies take advantage of the fact that most traits have a degree of genetic influence which can help predict the trait in a different genomic dataset. Transcriptome wide association study (TWAS) are the most prominent example. The first step is to determine the genetic variants that predict gene expression in one or more tissues. Next these variants are examined in GWAS data and summed to determine which genes are more or less expressed in cases compared to controls.
CLIC investigators, Drs. Tianzhong Yang, Saonli Basu, and Logan Spector have now published the first TWAS for the rare liver tumor hepatoblastoma (HB). With help from contributors from around the globe – Japan, Spain, Germany, Australia, Switzerland, United Kingdom, Italy, and multiple US locations – the team painstakingly assembled a dataset with genotypes from 5964 cases (4539 European, 1047 Latino, 378 AA) and 10:1 controls including publicly available data. Variants related to cross-tissue gene expression in the Genotype and Tissue Expression (GTEx) Project were then integrated into the GWAS.
Through TWAS analyses, the researchers found eight genes were predicted to have higher gene expression for children with hepatoblastoma than controls. Three of these genes were found in fetal liver hepatoblast cells and were differentially expressed in tumor and normal tissues in an independent Japanese HB study. Finally, they found that the genes were significantly enriched in the mitochondrial permeability transition pathway, potentially supporting the hypothesis that HB may be caused by oxidative damage.
The TWAS of HB is a great example of using integrative data to wring as much information out of a small GWAS as possible. CLIC plans to apply integrative methods like this to our data to “squeeze the juice” out of all our genomic resources.
For more information about the CLIC Genomics Project, please see our website: https://www.clic.ngo/introducing-the-clic-genomics-project/
Article Title: Multi-ancestry transcriptome-wide association study identifies candidate genes associated with hepatoblastoma.
Authors: Xie T, Sorenson JC, Spector LG, Pankratz N, Huang RS, Hiyama E, Poynter JN, Tomlinson GE, Armengol C, Kappler R, Scheurer ME, Roman E, Castellano A, Grotzer MA, Ziegler DS, Basu S, Marcotte EL, Yang T.
Published In: Cancer Epidemiol Biomarkers Prev. 2025 Jun 4. doi: 10.1158/1055-9965.EPI-24-1553. Epub ahead of print. PMID: 40465396.