An extreme athlete pushes his or her sport to the limit, striving to conquer the steepest slope, the tallest wave or the roughest terrain.
By those criteria, Dr. Hua Tang, a statistical expert in the Public Health Sciences Division, might be what you'd call an extreme genealogist. Not content merely to trace a family tree for a few generations, Tang has used DNA analysis to track humankind's genetic roots as far back as they can go. That turns out to be about 100,000 years ago in Africa, when many scientists believe that the ancestral "Adam and Eve" who gave rise to all living humans first emerged.
This comprehensive family tree reflects the migration of the first modern humans from Africa to other continents, where the geographically isolated populations eventually evolved unique characteristics that were passed on to their descendents.
Among the distinct traits of most interest to Tang are predispositions for certain diseases. She integrates her knowledge of human evolution and statistics to design powerful strategies to identify disease-susceptibility genes, a difficult research venture because of the many populations with mixed heritage that exist today.
Potential genetic tests
The outcome of this work could eventually lead to genetic tests to help doctors prevent serious diseases such as cancer or detect them at early onset, when they are most easily cured. Individuals found to be at high risk for particular conditions could be offered additional cancer screening or in some cases, drugs that prevent disease. In cases where a disease can't be prevented, some genes may also provide clues about which treatments will work best.
Distinguishing between genes that affect disease risk and those that do not has been a great challenge for scientists who study the causes of disease, said Tang, who joined PHS in 2002.
"There are many examples of where disease risk or disease outcome in response to treatment differs from subpopulation to subpopulation," she said. "But it can be difficult for scientists to determine whether the difference in risk is due to genetics, diet or environment or any other factors shared among a specific human subpopulation. This becomes especially challenging among population groups with mixed ancestry."
The reason for this dilemma stems from the type of study commonly used to identify genes that put individuals at higher risk of a certain disease. Known as a case-control study, it involves finding a group of individuals with a particular type of cancer (cases), for example, and a comparison group of individuals (controls) free of the disease.
Zeroing in on disease genes
Researchers then use techniques that allow them to comb through an individual's genetic blueprint to identify a part of the genome that is identical in many of the people with cancer but that differs in the controls. To maximize the likelihood that they will hone in on a gene truly connected to cancer risk, scientists strive to closely match the characteristics of each case — such as age, ethnicity, gender, smoking status, weight and other characteristics — with those of a control.
Yet even among seemingly well-matched cases and controls, scientists can't always conclude with certainty that the genes they identify are directly related to the disease being studied, Tang said.
For example, results of a case-control study might suggest that many breast-cancer cases share an identical gene that is rare among controls, leading researchers to conclude that having that gene puts an individual at higher-than-average risk for cancer. Yet those who possess the gene might have descended from similar ethnic roots — even without knowing it — that caused them to share genes that may have nothing to do with cancer risk.
"Case-control studies are powerful methods with many advantages and will continue to be used," Tang said. "But you can't always get accurate information by asking a person about their heritage. We're interested in developing methods that use genetic markers to evaluate ancestry, which will increase the power of case-control studies to find genes truly associated with disease risk."
Some of this work stems from research Tang conducted while a graduate student at Stanford University, where she earned a doctorate in statistics with a minor in genetics. In 2002, she published a paper describing a new statistical approach for determining the time to the most recent common ancestor of all living humans.
"For example, if you have a sister, the time to your most recent common ancestor is one generation, since you have the same parents," she said. "We wanted to know, how far back in time did the common ancestors to all of us live?"
Using an approach that involves extrapolating changes in DNA that occurred over time, Tang and colleagues estimated that the ancestral "Eve" lived about 200,000 years ago and the ancestral "Adam," about 100,000 years ago. The discrepancy between the sexes probably stems from tribal males at that time having multiple reproductive partners.
Statistical tools Tang developed for that project are helping her to apply the knowledge of human relatedness to overcome some of the challenges posed by case-control studies of disease. She points out that the very same genes that may give misleading results in these studies are useful as genetic markers of ethnicity.
Genetic markers are like barcodes along the chromosome that can differ slightly in their DNA sequence among individuals. The more closely related two people are, the more likely they are to have markers with identical DNA sequences. By identifying many markers in the genome and comparing them among individuals, researchers can estimate the relatedness between two individuals.
Using genetic markers of ancestry for case-control studies is especially important when studying population groups of mixed ancestry, such as Latinos and African-Americans. "This is why it's so important that studies include people from many ethnic backgrounds," Tang said. "Otherwise, we can't assess disease risk in all people."
Tang is putting this approach to the test through two collaborative studies that include participants of mixed ancestry. One, led by researchers at the University of California at San Francisco, seeks to identify genes linked to asthma in Latino Americans. The second, the Family Blood Pressure Program, will involve 18,000 participants from multiple ethnic groups to identify genes that influence high blood pressure.
Early detection biomarkers
Tang is also a member of the Early Detection Research Network, a National Cancer Institute-sponsored national effort to develop tests to detect the onset or risk of cancer using biomarkers-molecules such as proteins found in the blood, saliva or other body fluids that are released by cancers in their earliest stages. Under the direction of PHS investigator Dr. Ziding Feng, Fred Hutchinson houses the Data Management and Coordinating Center for the network. Dr. Nancy Kiviat, professor of pathology at the University of Washington and a PHS researcher, directs one of the consortium's 18 Biomarker Development Laboratories.
"In order to use biomarkers to screen individuals for cancer risk, it will be important to know the genetic variation among people, since some biomarkers might only turn out to be useful for certain populations," Tang said.
The ultimate achievement of such genetic studies would be truly individualized medical care, in which each person's risk of disease and response to treatments could be calculated based on his or her genetic profile, environmental exposures and diet. Tang said that although such comprehensive knowledge might be years off, important advances for specific illnesses could be made much sooner.
"It's more practical to try to understand how genetic variation affects risk at a coarser level-looking at one disease, in one population that is exposed to one risk factor," she said. "Multiple studies of this kind will help us accumulate a wealth of knowledge that will benefit human health."