Scientists have discovered a second code hiding within DNA. This second code contains information that changes how scientists read the instructions contained in DNA and interpret mutations to make sense of health and disease.

A research team led by Dr. John Stamatoyannopoulos, 91探花 associate professor of genome sciences and of medicine, made the discovery. The findings are reported in the Dec. 13 issue of Science.

Read the 聽 Also see commentary in Science, ”

The work is part of the Encyclopedia of DNA Elements Project, also known as ENCODE. The National Human Genome Research Institute funded the multi-year, international effort. ENCODE aims to discover where and how the directions for biological functions are stored in the human genome.

Since the genetic code was deciphered in the 1960s, scientists have assumed that it was used exclusively to write information about proteins. 91探花scientists were stunned to discover that genomes use the genetic code to write two separate languages. One describes how proteins are made, and the other instructs the cell on how genes are controlled. One language is written on top of the other, which is why the second language remained hidden for so long.

“For over 40 years we have assumed that DNA changes affecting the genetic code solely impact how proteins are made,” said Stamatoyannopoulos. “Now we know that this basic assumption about reading the human genome missed half of the picture. These new findings highlight that DNA is an incredibly powerful information storage device, which nature has fully exploited in unexpected ways.”

The genetic code uses a 64-letter alphabet called codons. The 91探花team discovered that some codons, which they called duons, can have two meanings, one related to protein sequence, and one related to gene control. These two meanings seem to have evolved in concert with each other. The gene control instructions appear to help stabilize certain beneficial features of proteins and how they are made.

The discovery of duons has major implications for how scientists and physicians interpret a patient’s genome and will open new doors to the diagnosis and treatment of disease.

“The fact that the genetic code can simultaneously write two kinds of information means that many DNA changes that appear to alter protein sequences may actually cause disease by disrupting gene control programs or even both mechanisms simultaneously,” said Stamatoyannopoulos.

Grants from the National Institutes of Health U54HG004592, U54HG007010, and UO1E51156 and National Institute of Diabetes and Digestive and Kidney Diseases FDK095678A funded the research.

In addition to Stamatoyannopoulos, the research team included Andrew B. Stergachis, Eric Haugen, Anthony Shafer, Wenqing Fu, Benjamin Vernot, Alex Reynolds, and Joshua M. Akey, all from the 91探花Department of Genome Sciences, Anthony Raubitschek of the 91探花Department of Immunology and Benaroya Research Institute, Steven Ziegler of Benaroya Research Institute, and Emily M. LeProust, formerly of Agilent Technologists and now with Twist Bioscience.

Stephanie H. Seiler heads the communications agency .

DNA analysis is unearthing the origins of the Minoans, who some 5,000 years ago established the first advanced Bronze Age civilization in present-day Crete. The findings suggest they arose from an ancestral Neolithic population that had arrived in the region about 4,000 years earlier.

The British archeologist Sir Arthur Evans in the early 1900鈥檚 named the Minoans after a legendary Greek king, Minos. Based on similarities between Minoan artifacts and those from Egypt and Libya, Evans proposed that the Minoan civilization founders migrated into the area from North Africa. Since then, other archaeologists have suggested that the Minoans may have come from other regions, possibly Turkey, the Balkans, or the Middle East.

Now, a team of researchers in the United States and Greece has used mitochondrial DNA analysis of Minoan skeletal remains to determine the likely ancestors of these ancient people.

Mitochondria, the energy powerhouses of cells, contain their own DNA, or genetic code. Because mitochondrial DNA is passed down from mothers to their children via the human egg, it contains information about maternal ancestry.

Results published May 14 in Nature Communications suggest that the Minoan civilization arose from the population already living in Bronze Age Crete. The findings indicate that these people probably were descendents of the first humans to reach Crete about 9,000 years ago, and that they have the greatest genetic similarity with modern European populations.

Read the .

Dr. George Stamatoyannopoulos, 91探花 professor of medicine and genome sciences, is the paper鈥檚 senior author. He believes that the data highlight the importance of DNA analysis as a tool for understanding human history.

鈥淎bout 9,000 years ago,鈥� he noted, 鈥渢here was an extensive migration of Neolithic humans from the regions of Anatolia that today comprise parts of Turkey and the Middle East. At the same time, the first Neolithic inhabitants reached Crete.鈥�

鈥淥ur mitochondrial DNA analysis shows that the Minoan鈥檚 strongest genetic relationships are with these Neolithic humans, as well as with ancient and modern Europeans,鈥� he explained.

鈥淭hese results suggest the Minoan civilization arose 5,000 years ago in Crete from an ancestral Neolithic population that had arrived in the region about 4,000 years earlier,鈥� he said. 鈥淥ur data suggest that the Neolithic population that gave rise to the Minoans also migrated into Europe and gave rise to modern European peoples.鈥�

Stamatoyannopoulos, who directs the 91探花Markey Molecular Medicine Center and who formerly headed the 91探花Division of Medical Genetics in the Department of Medicine, added, 鈥淕enetic analyses are playing in increasingly important role and predicting and protecting human health. Our study underscores the importance of DNA not only in helping us to have healthier futures, but also to understand our past.鈥�

Stamatoyannopoulos and his research team analyzed samples from 37 skeletons found in a cave in Crete鈥檚 Lassithi plateau and compared them with mitochondrial DNA sequences from 135 modern and ancient human populations. The Minoan samples revealed 21 distinct mitochondrial DNA variations, of which six were unique to the Minoans and 15 were shared with modern and ancient populations. None of the Minoans carried mitochondrial DNA variations characteristic of African populations.

Further analysis showed that the Minoans were only distantly related to Egyptian, Libyan, and other North African populations. The Minoan shared the greatest percentage of their mitochondrial DNA variation with European populations, especially those in Northern and Western Europe.

When plotted geographically, shared Minoan mitochondrial DNA variation was lowest in North Africa and increased progressively across the Middle East, Caucasus, Mediterranean islands, Southern Europe, and mainland Europe. The highest percentage of shared Minoan mitochondrial DNA variation was found with Neolithic populations from Southern Europe.

The analysis also showed a high degree of sharing with the current population of the Lassithi plateau and Greece. In fact, the maternal genetic information passed down through many generations of mitochondria is still present in modern-day residents of the Lassithi plateau.

Co-authors of the study are Jeffery R. Hughey of Hartnell College; Peristera Paschou of Democritus University of Thrace; Petros Drineas of the Rensselaer Polytechnic Institute; Manolis Michalodimitrakis of the University of Crete; and Donald Mastropaolo, Dimitra M. Lotakis, Patrick A. Navas, and John A. Stamatoyannopoulos of the 91探花. The study was partially supported by a grant from the National Institutes of Health (5T32 GM007454), as well as from private funding.

###

Dr. John A. Stamatoyannopoulos, associate professor of genome sciences and medicine at the UW, director of the 91探花ENCODE center, and a senior author on seven ENCODE-related papers, explains why understanding how the human genome functions is important to progress in genomic medicine:

鈥淭he first phase of the human genome project provided the primary genome sequence, and a basic catalog of genes, which occupy only 2 percent of the genome.聽 Every cell in the body has the same genes, but different kinds of cells, such as liver or heart, switch on different combinations of genes.聽 When cells become unhealthy, these combinations change.聽 Understanding how genes turn on and off is therefore vital to deciphering their role in both normal health and disease.聽 The instructions for how genes are controlled are contained in small DNA ‘switches’ that are scattered around the 98 percent of the genome that does not contain genes.聽 Mapping and decoding these instructions is a central mission of the ENCODE project, and the focus of work at the 91探花ENCODE center.聽 Data generated in this project so far have already shown, for example, that common DNA variations in the gene-controlling switches can affect the risk of developing different common diseases. This finding, together with the emerging wealth of information about the basic mechanisms of gene control, is opening new vistas on preventing, diagnosing, and treating disease.鈥�

Stamatoyannopoulos is also author of a headlining article for the ENCODE special issue of Genome Research that gives an overview and perspective on the project, its accomplishments, their significance, and prospects for the future. He participated in a national and international聽 press briefing this morning organized by the National Human Research Genome Institute. He discussed the importance of the findings to medical genomics.

At the London Museum of Science, where the Project ENCODE news announcement was made at noon GMT, the accomplishments of the hundreds of researchers and many countries involved in the effort were celebrated. Silk banners imprinted with ENCODE data dropped from the ceiling of the museum聽 for a performance by aerial dancers. See of the Dance of DNA.

Many 91探花researchers contributed to ENCODE-related research.聽 Major discoveries include:

The first detailed maps of regulatory DNA switches that make up the genome’s ‘operating system.

Researchers located millions of DNA ‘switches’ that dictate how, when, and where in the body different genes turn on and off.聽 These switches, or regulatory DNA, contain small chains of DNA ‘words’ that make up docking sites for proteins involved in gene control.聽 Often these switches are far away from the genes that they control. Of the millions of regulatory DNA regions, only a small fraction, around 200,000, are active in any given cell type.聽 This fraction is almost unique to each type of cell, a sort of molecular bar code of its identity.聽 The regulatory ‘program’ of most genes has more than a dozen switches. Nature paper: The accessible chromatin landscape of the human genome.

The first extensive map of regulatory protein docking sites on the human genome reveals the dictionary of DNA words that comprise the genome’s programming language.

To find the DNA words recognized by regulatory proteins, researchers employed a simple, powerful trick to study all the proteins at once.聽 Instead of trying to see proteins directly, they looked for their footprints on the DNA. They discovered that over 90 percent of the protein docking sites were actually slight variants of about 680 different DNA words — a dictionary of the genome’s programming language. Nature paper: An expansive human regulatory lexicon encoded in transcription factor footprints.

A comprehensive wiring diagram provides insights into how cells ‘think’

The genome senses and responds to signals received from other parts of the cell and from the environment by changing the activity of regulatory proteins. Scientists mapped all of the connections between regulatory protein genes to create a central wiring diagram for the cell.聽 Using powerful computers, they created, in a matter of weeks, wiring diagrams o how 475 regulatory protein genes were connected to each other, and how those connections changed across 41 different types of human cells.聽 Conventional methods would have required nearly 20,000 different experiments, taking several years to complete and costing over one hundred million dollars. Even though individual connections between regulatory proteins differed among cell types, the overall connection was nearly the same in all cell types.聽 When compared to the best-studied biological network — the map of all connections between neurons in the worm brain, created by Nobel Prize winner Sydney Brenner 鈥� the layout is almost identical.聽 Nature seems to have settled on an ideal ‘brain-like’ architecture to process complex biological information; this plan can be found in the genomic wiring of every living cell.聽 Cell paper: Circuitry and dynamics of human transcription factor regulatory networks.

Unlocking disease information hidden in the genome’s control circuitry.

Hundreds of studies have attempted to map the genes causing common diseases and physical traits.聽 Frustratingly, most of these studies have pointed to regions of the genome that do not contain gene sequences that make protein. Researchers set out to chart a global map of the relationship between disease-associated genetic changes and the gene-controlling switches scattered around the genome, With support from National Institutes of Health鈥檚 Common Fund, researchers collected regulatory DNA maps from 349 tissue samples covering all major organ systems in adults and stages of human development.聽 Using powerful computers, the researchers crossed these maps with data from genetic studies of over 400 common diseases and clinical traits.聽 Instead of isolated instances, they found that most disease-associated genetic changes occurred within gene-regulating switches, often located far away from the genes they control.聽 Most changes affected circuits active during early human development, when body tissues are most vulnerable. Extensive blueprints of control circuitry revealed previously hidden connections between diverse diseases, may explain common clinical features, and will open new avenues for developing diagnostics and treatments.聽 Science (cover feature):聽 Systematic localization of common disease-associated variation in regulatory DNA.

Differences in regulatory DNA between people and human populations, and evolutionary changes from natural selection

In comparing genomes from individuals from several parts of the world, a team led by Benjamin Vernot, Joshua Akey, and Stamatoyannopolous found that changes affecting gene control regions are very frequent. In the average individual these changes dwarf those found in DNA that encodes proteins.聽 By performing genome-wide scans for areas of recent evolutionary change, scientist discovered evidence that hundreds regulatory DNA regions had been targeted by natural selection, presumably because of their roles in biological pathways important for human survival, such as skin pigmentation and fat storage. Genome Research: Personal and population genomics of human regulatory variation.

New insights into the genome’s ‘master weaver’

Not all regulatory proteins are created equal: CTCF has earned the title of the genome’s ‘master weaver’ because it not only controls genes but determines how DNA is wound up within the cell nucleus.聽 Prior research had suggested that the sites along the genome to which CTCF liked to dock were nearly the same in every cell.聽 By CTCF binding across many cell types, 91探花researchers have found patterns varied among different cells, and between normal cells and cells that keep growing indefinitely, such as cancer cells. Many changes between cell types were accompanied by chemical changes in DNA known as methylation, which has been linked with aging.聽 Genome Research: Widespread plasticity in CTCF occupancy linked to DNA methylation.

Teaching computers to find patterns in ENCODE “big data”

91探花researchers in Genome Sciences and Electrical Engineering, led by William Noble and Michael M. Hoffman and have applied machine learning techniques to find patterns in the structure of human DNA and associated biomolecules. Machine learning involves the use of computer programs that can learn to classify big data sets into human-interpretable categories. The team designed a computer program and trained it to examine and characterize data on the location of chromatin modifications. Chromatin is the three-dimensional structure that DNA forms as it wraps around bead-like protein structures known as nucleosomes. The team used the software to identify patterns associated with genes and other DNA elements important in the regulation of gene activity.

Without these switches, called regulatory DNA, genes are inert. Researchers around the world have been focused on identifying regulatory DNA to understand how the genome works. Using a new technology developed with funding from the National Human Genome Research Institute’s ENCODE (ENCyclopedia Of DNA Elements) project, 91探花researchers created the first detailed maps of where regulatory DNA is located within hundreds of different kinds of living cells. They also compiled a dictionary of the instructions written within regulatory DNA — the genome’s programming language.

The findings are reported in two papers appearing in the Sept. 5 online issue of Nature.

“These breakthrough studies provide the first extensive maps of the DNA switches that control human genes,” said Dr. John A. Stamatoyannopoulos, associate professor of genome sciences and medicine at the 91探花, and senior author on both papers. “This information is vital to understanding how the body makes different kinds of cells, and how normal gene circuitry gets rewired in disease. We are now able to read the living human genome at an unprecedented level of detail, and to begin to make sense of the complex instruction set that ultimately influences a wide range of human biology.”

Here are the key results:

1) The first detailed maps of regulatory DNA switches that make up the genome’s ‘operating system’.

The instructions within regulatory DNA are inscribed in small DNA ‘words’ that function as the docking sites for special proteins involved in gene control. In many cases, these switches are located far away from the genes that they control. To map the regulatory DNA regions, the researchers harnessed a special molecular probe — an enzyme called DNaseI — that snips the genome’s DNA backbone. Under the right conditions, these snips occur precisely where proteins are docked at regulatory DNA. By treating cells with DNase I and analyzing the patterns of snipped DNA sequences using massively parallel sequencing technology and powerful computers, the researchers were able to create comprehensive maps of all the regulatory DNA in hundreds of different cell and tissue types. They found that of the 2.89 million regulatory DNA regions they mapped, only a small fraction — around 200,000 — were active in any given cell type. This fraction is almost totally unique to each type of cell and becomes a sort of molecular bar code of the cell’s identity. The researchers also developed a method for linking regulatory DNA to the genes it controls. The results of these analyses show that the regulatory ‘program’ of most genes is made up of more than a dozen switches. Together, these findings greatly expand the understanding of how genes are controlled and how that control may differ between normal and diseased cells.

2) The first extensive map of regulatory protein docking sites on the human genome reveals the dictionary of DNA words comprise the genome’s programming language.

The instructions for turning genes on and off are written in DNA switches called regulatory DNA. These switches are scattered throughout the non-gene regions of the human genome. Having mapped the locations of the regulatory DNA switches, 91探花researchers wanted to know what made them tick. These regions contain small chains of DNA ‘words’ that make up docking sites for special regulatory proteins involved in gene control. The human genome contains hundreds of genes that make such proteins.

However, current technologies only allow such proteins to be studied one at a time. They also lack the accuracy to resolve the DNA letters to which the proteins dock. As a result, most of the actual DNA words recognized by regulatory proteins in living cells were unknown. To find them, the researchers employed a simple, powerful trick that enabled them to study all the proteins at once.

Instead of trying to see proteins directly, they looked for their shadows or ‘footprints’ on the DNA. To accomplish this, they again turned to the DNaseI enzyme that snips the DNA backbone within regulatory DNA. Prior work had shown that DNaseI likes to snip DNA next to regulatory protein docking sites, but not within the docking site itself. By using next-generation DNA sequencing technology, the researchers analyzed hundreds of millions of DNA backbone breaks made when cells were treated with DNaseI. They then used a powerful computer to resolve millions of protein footprints. In total, they identified 8.4 million such footprints along the genome, some of which were detected in many cell types. Next, they compiled all of the short DNA sequences to which the proteins were docked. They analyzed them using a software algorithm that required hundreds of microprocessors working simultaneously. This revealed that more than 90 percent of the protein docking sites were actually slight variants of 683 different DNA words — essentially a dictionary of the genome’s programming language.

“These findings significantly advance the understanding of how the instructions for controlling genes are written and organized throughout the genome, and how combinations of different instruction sets function together to control genes, often at great distance along the genome,” Stamatoyannopoulos said. “The broad spectrum of cell and tissue types included in these analyses provide an incredibly rich resource that can be mined immediately by researchers around the world to illuminate how the genes they are studying are controlled.”

The scientists determined that genes are connected in a complex web. In this web, regulatory DNA regions typically control one or at most a few genes, but genes receive inputs from large numbers of regulatory regions. The researchers also found evidence for a combinatorial code that helps match regulatory DNA with the right genes. Another key finding was that the regulatory DNA controlling genes involved in cancer and other types of ‘immortal’ cells that can keep on growing indefinitely appears to acquire mutations at a different rate than other kinds of regulatory DNA. This result points to a previously unknown link between genome function and patterns of DNA variation in individual human genomes. The finding may have implications for understanding susceptibility to cancer.

The findings reported in these papers are expanded upon in two related papers to be published simultaneously in the journals Science and Cell. In the Science paper, 91探花researchers further expanded the regulatory DNA maps, and compared them with genetic maps of human disease. Their studies revealed that most DNA variants associated with specific human diseases or clinical traits are located in regulatory DNA rather than in gene sequences. In the Cell paper, the researchers describe using the detailed information on regulatory protein docking sites to create a comprehensive map of how those proteins are wired.

Researchers at the 91探花 have determined that the majority of genetic changes associated with more than 400 common diseases and clinical traits affect the genome’s regulatory circuitry. These are the regions of DNA that contain instructions dictating when and where genes are switched on or off. Most of these changes affect circuits that are active during early human development, when body tissues are most vulnerable.

By creating extensive blueprints of the control circuitry, the research also exposed previously hidden connections between different diseases. These connections may explain common clinical features, as well as offer a new approach for pinpointing the specific types of cells and tissues that either cause or are most affected by a particular disease. The findings provide a major paradigm shift for understanding the genetic causes of disease, and open new avenues for development of diagnostics and treatments. The findings appear in the Sept. 5 online issue of Science.

“Genes occupy only a tiny fraction of the genome, and most efforts to map the genetic causes of disease were frustrated by signals that pointed away from genes. Now we know that these efforts were not in vain, and that the signals were in fact pointing to the genome’s ‘operating system’ — the instructions for which are hidden in millions of locations around the genome,” said Dr. John A. Stamatoyannopoulos, associate professor of genome sciences and medicine at the UW. “The findings provide a new lens through which to view the role of genetics and genome function in disease.”

The human genome’s control circuitry is encoded in millions of regulatory regions — short DNA sequences that are scattered throughout the 98 percent of the genome that does not specify the protein product of a gene. Specialized proteins, called regulatory factors, recognize specific DNA sequences in these regulatory regions, thereby creating switches that turn genes on and off. In many cases, these switches are located far away from the genes that they control. These distances have made it difficult to determine the relationship between specific switches and genes.

The researchers used a special molecular probe called a nuclease to detect all of the regulatory regions active in each cell type they studied. The specific nuclease they used — called DNase I — snips the genome where regulatory factors are bound to DNA. By treating cells with DNase I and analyzing the pattern of snipped DNA sequences using massively parallel sequencing technology and high-performance computers, the researchers were able to create comprehensive maps of all the regulatory DNA in many different types of cells. These maps were then analyzed with advanced software algorithms to sort through the data and expose previously hidden connections between disease-associated genetic variation and specific regulatory regions.

The regulatory mapping and analysis was conducted on 349 cell and tissue samples. These included samples from all major organs as well as 233 tissue samples from different stages of early human development. In total, nearly 4 million distinct regulatory regions were discovered, though only about 200,000 of these were ‘on’ in any particular cell type.

To make a connection with common diseases and clinical traits, the researchers analyzed genetic variants that had been strongly associated with diseases and traits through so-called genome-wide association studies, which compare genetic information between groups of people with or without a particular disease or trait. During the past decade, hundreds of genome-wide association studies involving hundreds of thousands of patients worldwide have been performed for over 400 diseases and traits. Nearly 95 percent of the time, these studies flagged genetic variants that were located outside of gene protein-coding regions. Comparison of these data with the regulatory DNA blueprints yielded several key findings:

• 76 percent of disease-associated variants in non-gene regions are actually located within or are tightly linked to regulatory DNA. This suggests that many diseases result from changes in when, where, and how genes are turned on rather than changes to the gene itself.
• 88 percent of the regulatory regions that contained disease-associated DNA variants were active in early human development fetal development. Because many of these variants are associated with common diseases that occur in adults, the finding indicates that factors influencing the genome’s regulatory circuitry early in development may impact the risk of developing particular diseases later in life.
• DNA changes associated with specific diseases tend to occur in the specific short DNA codes recognized by regulatory proteins involved in physiological processes related to the disease or the organs or cells affected by the disease. For example, DNA variants associated with diabetes tend to occur in the codes recognized by regulatory proteins that control various aspects of sugar metabolism and insulin secretion. Similarly, variants associated with immune system disorders, such as multiple sclerosis, asthma, or lupus, are found in specific recognition codes for proteins that regulate immune system function.
• Many seemingly unrelated diseases share common regulatory circuitry, including diseases that affect the immune system, different types of cancers, and a range of neuropsychiatric disorders.

The study also revealed a wealth of additional connections between genetic variants and disease that had been lurking within existing genome-wide association studies data. Viewing these data through the lens of regulatory DNA exposed thousands of variants that were highly selectively localized within regulatory DNA of disease-specific cell types. These variants had previously been ignored because the stringent selection criteria used in earlier studies did not take regulatory regions into account.

Another surprising finding was that the regulatory circuitry blueprints could be used to pinpoint cell types that play a role in specific diseases — without requiring any prior knowledge about how the disease worked. For example, genetic variants associated with Crohn’s disease (a common type of inflammatory bowel disease) were found to be concentrated in the regulatory regions mapped in two specific subsets of immune cells — the same cell types that took decades of prior research to be linked with development of tCrohn’s disease. Applying this approach systematically will enable researchers to identify cell types not previously known to play a role in a particular disease, expanding our understanding of the disease process and potentially leading to new therapies.

The study was supported in part by the National Institute’s of Heath (NIH) Common Fund Roadmap Epigenomics Program (grant number U01ES017156), the National Human Genome Research Institute’s ENCODE Project (U54HG004592), the National Institute of Child Health and Human Development (R24HD000836-47), the National Institute of Diabetes Digestive and Kidney Diseases (P30 DK056465), and the National Heart, Lung, and Blood Institute (R01HL088456). The NIH Common Fund supports a series of exceptionally high impact research programs that are broadly relevant to health and disease. Common Fund programs are designed to overcome major research barriers and pursue emerging opportunities for the benefit of the biomedical research community at large. The research products of Common Fund programs are expected to catalyze disease-specific research supported by the NIH Institutes and Centers.