Epigenome mapped in detail across human body

A human body consists of trillions of cells, each of which has the same instruction manual within, a sequence of DNA three billion letters long. There are chemical compounds, including proteins, which modify how the instruction manual is interpreted, for example by attaching to DNA and turning a gene on or off. The chemical compounds which tell the genome what to do are collectively known as the epigenome.

The epigenome is the reason why a neuronal cell in the brain and a muscle cell in the heart have completely different shapes and sizes despite having the same instruction manual. As cells become specialised, their epigenome starts to change and diverge from the epigenome of other cell types. When specialised cells divide, much of their epigenome is passed on to the next generation of cells, resulting in entire tissues having a unique set of genomic modifications, also known as epigenomic “marks”.

Changes in the epigenome can cause, or result from, disease, so charting the locations and understanding the function of epigenomic marks is central to human biology. Creating epigenomic maps of each tissue is vital for the future of personalised medicine, as it can one day enable doctors to determine an individual’s health and tailor a patient’s response to medication.

An international research collaboration has addressed this challenge by cataloguing the human epigenome in unprecedented detail, charting the whole collection of epigenomic marks in more than 25 different types of tissues across the human body. The findings, published today in the journal Cell, provide the most comprehensive map of the human epigenome to date.

“Each individual is unique and has a degree of variation, meaning that my liver’s epigenome will differ from yours even though it’s the same tissue. We expect to see even greater changes in a state of disease, and this new resource will help us measure these differences in order to understand the mechanisms of disease,” explains Dr. Roderic Guigó, senior co-author of the study and researcher at the Centre for Genomic Regulation in Barcelona.

The study generated the epigenome dataset by sequencing the genomes of four individual donors and studying the activity of all of the genes and their regulatory regions across 30 different types of tissues, including data from hard-to-obtain tissues such as the lung. Known as EN-TEx, the personal epigenomes generated by the project can be used as a reference and combined with other human genome annotations to discern whether a genetic variant in a particular individual contributes to health or disease.

One of the unique strengths of EN-TEx is that it’s the first resource of its kind to include information from both copies of the chromosomes from each individual. It means that, for the first time, researchers will be able to discern the impact of maternal or paternally-inherited genetic variants on human biology.

"An individual has between 4 and 4.5 million mutations, but it's hard to know which ones are harmful. This resource allows us for the first time to know which mutations an individual has inherited from mom and dad, and understand which of them may potentially have an impact on their health,” explains Dr. Beatrice Borsari, who carried out the work during her PhD at the Centre for Genomic Regulation in Barcelona and is currently a postdoctoral researcher at Yale University in the United States.

EN-TEx overcomes one of the limitations of the original reference genome published in 2003. This was assembled using only one copy of each chromosome, meaning that genetic variants which are specific to one of the two copies of a chromosome – known as allele-specific (AS) variants – were missed or incorrectly represented. Successive large-scale initiatives such as the ENCODE or GTEx projects, which used the reference genome as the foundation for their findings, also failed to include these types of genetic variants. The research team used the resource to identify and locate more than one million AS variants in the genome, significantly more than what was previously known.

One of the applications of the resource is predicting the behaviour of DNA sequences that control how genes behave, also known as expression quantitative trait loci or eQTLs. Researchers leveraged the data in EN-TEx and used machine learning to build a tool which can transfer eQTLs from one tissue to another. In other words, it is now possible to detect how certain DNA sequences influence the behaviour of genes in hard-to-obtain tissues like the lung by measuring them from a blood sample instead, knowledge that can help develop new therapies and treatments for tissues that are hard to study, such as the heart or the lung.

Another important application of EN-TEx is being able to predict the behaviour of a class of proteins that modify the genome known as transcription factors. Mutations or changes in the regulation of transcription factors can lead to a wide range of diseases, including cancer, metabolic disorders, and immune disorders. As a result, transcription factors are an important target for the development of new therapeutics.

The team used EN-TEx to develop a deep learning model that can predict whether a variant can disrupt the binding site of transcription factors. The model revealed that researchers need to look beyond just the binding site itself and also consider the area around the site. The team found that the key to whether a binding site would be disrupted was the presence of nearby binding sequences for other regulatory factors.

“Think of regulatory factors as the legs of the Lunar Module,” says Mark Gerstein, PhD, professor of biomedical informatics at Yale University in a Yale School of Medicine news release. “If it has four legs and one leg doesn’t work, the three other legs can anchor the defective leg. Similarly, the anchoring of other regulatory factors might stabilize the disrupted binding site and make it less sensitive to variants.”

One limitation of the resource is that only four people of European descent are profiled. The team would like to eventually enlarge their study to encompass hundreds of individuals with more diverse ancestries.

The study was funded by the United States of America’s National Human Genome Research Institute (NHGRI) and led by researchers from Yale University, Harvard University, the Massachusetts Institute of Technology (MIT), Johns Hopkins University and Cold Spring Harbor Laboratory in the United States, as well as the Centre for Genomic Regulation in Barcelona.

Collaborating institutions including Baylor College of Medicine; California Institute of Technology; the Dana-Farber Cancer Institute; the European Bioinformatics Institute; HudsonAlpha Institute for Biotechnology; New York Institute of Technology; Stanford University; University of California, Irvine; University of California, San Diego; University of Hong Kong; University of Massachusetts Medical School; University of Toronto, Canada; and University of Washington, Seattle.

The Centre for Genomic Regulation’s role in the international collaboration includes participating in the design of the project and contributions to many areas of the subsequent research; developing models to transfer eQTLs among tissues, as well as improving the accuracy of analyzing data obtained from individual genomes rather than the reference genome.

"Catalonia has played a role in large-scale genomics initiatives over the last twenty years, including the Human Genome Project. Genomics is one of the fields which will have the greatest impact on people's lives in the 21st century. We have shown we can play an important role in these projects, but we also need to take charge and launch our own large-scale projects too. We are in a position to do so. For that, we need more resources, but above all it forward-thinking and ambition," concludes Dr. Guigó.