American Association for Cancer Research

Focus on Science January 2014

Elaine R. Mardis, Ph.D.Elaine R. Mardis, Ph.D.
Co-director, The Genome Institute
Professor, Washington University

Dr. Mardis joined The Genome Institute in 1993. As director of technology development she has helped create methods and automation pipelines for sequencing the human genome. She was named co-director of the Genome Institute in 2002.

Dr. Mardis has research interests in the application of DNA sequencing to characterize cancer genomes. She also is interested in facilitating the translation of basic science discoveries about human disease into the clinical setting.

"Cancer genomics efforts in my laboratory have focused on two general areas: integrated cancer genomics discovery and translating cancer genomics to clinical care. Over the past few years, while our understanding of the genomic landscape of cancer has dramatically improved due to next-generation sequencing, we have continued to expand the characterization of cancer cell nucleic acids by studying the transcriptome (expressed RNA) and the methylome (changes to the methylation of DNA when comparing tumor and matched normal DNA from the same tissue). Integrating these different types of data provides an enhanced view of the somatic alterations that are unique to the cancer genome and gives us a richer understanding of all the changes that impact tumor biology. The size of the genome, the complexity of RNA expression and the resolution required to characterize methylation changes increase the difficulty of these efforts, which we approach from both the technology and computational biology angles. Our laboratory has been involved in local- and network-based collaborative efforts toward basic discovery in cancer across multiple tumor types.

"One of our network-based collaborations is the Pediatric Cancer Genome Project (PCGP) with St. Jude Children’s Research Hospital. This effort, during its initial three years, characterized the somatic alterations for 750 pediatric cancer cases (matched tumor and normal tissues) across multiple tumor types by whole genome sequencing and data analysis [1]. The resulting discoveries have been published in several manuscripts, but importantly all of the data have been made publicly available for other researchers to access and continue the discovery process. Ultimately, our efforts and those of others will enhance our understanding of pediatric cancer genomes, leading to the identification of new diagnostics, therapeutics, and prognostic markers. Recently, the PCGP was extended for two years and will focus on integrated data generation and analysis for 100 cases each of acute lymphoblastic leukemia, medulloblastoma and neuroblastoma. Here, we will produce whole genome, exome, RNA, and whole genome methylation sequencing for each tumor, thereby creating a comprehensive catalog of somatic alterations for these three major pediatric tumor types.

"A similar network collaborative effort in adult cancer is The Cancer Genome Atlas (TCGA) project that is jointly funded by the National Human Genome Research Institute and the National Cancer Institute. In particular, our laboratory has produced the sequencing data and mutational analyses for two recently published studies, on breast cancer and endometrial cancer [2, 3]. I was fortunate to have co-chaired the analysis working group for TCGA’s endometrial cancer project with Dr. Douglas Levine and the Nature manuscript that resulted has shed important new light on endometrial cancer genomics, including frequently mutated genes, potential new biomarkers and their associated prognoses. The overarching aim of basic discovery efforts in genomics is to provide fundamental information toward improving clinical cancer care, where our efforts also have focused.

"One translational project that completed recently was an effort to characterize and compare, by whole genome sequencing, transcriptome sequencing and proteomic/phosphoproteomic analysis (RPPA), 13 primary human breast tumors to their xenografted counterparts, propagated in NSG mice [4]. Our resulting manuscript described the result that across DNA, RNA and protein assays, the xenografted tumor faithfully represents the primary tumor from which it was derived. This foundational knowledge is critically important as we and others would like to begin testing the correlation between mutated genes and therapies that directly bind or “target” their protein products. In demonstrating that xenografted tumors are representative of the biology of their human progenitors, we established that drug-gene assay results largely will faithfully represent the response of the human tumor to each specific therapy tested. In addition to this foundational discovery, we further provided additional evidence that alterations to the estrogen receptor gene (ESR1) appear to correlate to endocrine therapy-resistance in late relapse ER+ breast cancer. This result was recently further supported by additional manuscripts showing mutation in ESR1 yields this phenotype. The resulting implications for patient care are now being explored.

"Our Drug-Gene Interaction database, DGIdb was recently published and this resource continues to evolve [5]. DGIdb is designed to be a clearinghouse for readily searching and linking information that is typically difficult to integrate, namely identifying drugs and the gene product(s) they target. In general, human genes are referred to by multiple different names or abbreviations, and drugs have a multitude of names and can target multiple proteins. We designed DGIdb to be a decision support tool for clinicians to help them interpret the genes often found mutated in human diseases (especially cancer); linking genes to the drugs designed to target their mutated protein products, and to any clinical trials that are actively ongoing for the indicated drugs. A simple search interface permits these associations to be discovered, then allows the user to link out to supporting information housed in other databases that can help to evaluate the drugs being considered, based on published research, clinical trials, or other information. As is true for all our clinical translation efforts, we hope this resource will facilitate the application of targeted therapies to clinical cancer care."


1. Downing, J.R., et al., The Pediatric Cancer Genome Project. Nat Genet, 2012. 44(6): p. 619-22.
2. Comprehensive molecular portraits of human breast tumours. Nature, 2012. 490(7418): p. 61-70.
3. Kandoth, C., et al., Integrated genomic characterization of endometrial carcinoma. Nature, 2013. 497(7447): p. 67-73.
4. Li, S., et al., Endocrine-therapy-resistant ESR1 variants revealed by genomic characterization of breast-cancer-derived xenografts. Cell Rep, 2013. 4(6): p. 1116-30.
5. Griffith, M., et al., DGIdb: Mining the druggable genome. Nat Methods, 2013. 10(12): p. 1209-10.