Risks and Rewards of Big-Data in Epigenomics Research: An Interview with Melanie Ehrlich

Abstract

Melanie Ehrlich, PhD, is a professor in the Tulane Cancer Center, the Tulane Center for Medical Bioinformatics and Genomics and the Hayward Human Genetics Program at Tulane Medical School, New Orleans, LA. She obtained her PhD in molecular biology in 1971 from the State University of New York at Stony Brook and completed postdoctoral research at Albert Einstein College of Medicine in 1972. She has been working on various aspects of epigenetics, starting with DNA methylation, since 1973. Her group made many first findings about DNA methylation (see below). For example, in 1982 and 1983, in collaboration with Charles Gehrke at the University of Missouri, she was the first to report tissue-specific and cancer-specific differences in overall DNA methylation in humans. In 1985, Xian-Yang Zhang and Richard Wang in her lab discovered a class of human DNA sequences specifically hypomethylated in sperm. In 1998, her group was the first to describe extensive losses of DNA methylation in pericentromeric and centromeric DNA repeats in human cancer. Her lab's many publications on the prevalence of both DNA hypermethylation and hypomethylation in the same cancers brought needed balance to our understanding of the epigenetics of cancer and to its clinical implications [ 1 ]. Besides working on cancer epigenetics, her research group has helped elucidate cytogenetic and gene expression abnormalities in the immunodeficiency, centromeric and facial anomalies (ICF) syndrome, a rare recessive disease often caused by mutations in DNMT3B. Her group also studied the epigenetics and transcriptomics of facioscapulohumeral muscular dystrophy (FSHD), whose disease locus is a tandem 3.3-kb repeat at subtelomeric 4q (that happens to be hypomethylated in ICF DNA [ 2 ]). Her study of FSHD has taken her in the direction of muscle (skeletal muscle, heart and aorta) epigenetics [ 3–6 ]. Recently, she has led research that applies epigenetics much more rigorously than usual to the evaluation of genetic variants from genome-wide association studies (GWAS) of osteoporosis and obesity. In continued collaboration with Sriharsa Pradhan at New England Biolabs and Michelle Lacey at Tulane University, she has compared 5-hydroxymethylcytosine and 5-methylcytosine clustering in various human tissues [ 7 ] and is studying myoblast methylomes that they generated by a new high-resolution enzymatic technique (enzymatic methyl-seq).

Big Data in Laboratory Medicine—FAIR Quality for AI?

“…An obvious field for Big Data technologies in laboratory medicine are “-omics” applications [ 59 , 60 , 61 ]. These have been developed for nucleic acid-based techniques as, e.g., genomics [ 62 , 63 ], transcriptomics [ 64 ], and epigenomics [ 65 ], as well as for mass spectrometry-based methodologies such as proteomics [ 66 , 67 ], metabolomics [ 68 , 69 ], lipidomics [ 70 ], and others. The particular challenges in this field include connecting the analysis systems to the corresponding data lakes—it is no longer possible to work with traditional database technologies and new approaches, for example, hadoop [ 71 ] become necessary.…”

Big Data in Laboratory Medicine – FAIR Quality for AI?

“…An obvious field for Big Data technologies in laboratory medicine are "-omics" applications [57][58][59]: These have been developed for nucleic acid-based techniques as e.g. genomics [60,61], transcriptomics [62], and epigenomics [63], as well as for mass spectrometry-based methodologies such as proteomics [64,65] , metabolomics [66,67], lipidomics [68] and others. The particular challenges in this field include connecting the analysis systems to the corresponding data lakes -it is no longer possible to work with traditional database technologies and new approaches for example hadoop [69] become necessary.…”