The investigators, Dr. Melodie Winawer, assistant professor of neurology, Dr. Daniel Rabinowitz, professor of statistics, and Dr. Ruth Ottman, professor of epidemiology and deputy director for research at the Gertrude H. Sergievsky Center, most recently used their method to find evidence for genetic influences on two common types of epilepsy. The study was featured in the highlights section of the Dec. 9 issue of Neurology.
"Our results should help group families in a way that makes it easier to find genetic causes," says Dr. Winawer, the first author on the paper. "Since epilepsy is really a complex set of diseases with varied symptoms and multiple genetic and nongenetic causes, it is important to make rational subdivisions when looking for susceptibility genes."
The researchers examined whether members of families tend to have the same types of epilepsy, which might indicate a genetic influence on type. The method, concordance analysis, examines how often all individuals with epilepsy in a family have the same type of epilepsy and then determines whether the distribution of types in families differs from that expected by chance. Their method addresses a flaw in previous studies, which examined epilepsy types in families without testing whether the observations differed from chance expectations.
The researchers realized this weakness about five years ago and began developing a new method to determine which symptoms or types of epilepsy are controlled by genes. Their statistical method is based on a permutation test, which in its simplest form is like pulling marbles out of a bag that contains black and white marbles. The computer test simulates repetitive sampling to calculate how often, by chance, all members in a family have the same type of epilepsy, which would be comparable to drawing only black marbles or only white ones from the bag. Then the observed number of families in which everyone has the same type of epilepsy is compared with the number expected by chance. Any clinical feature of epilepsy can be analyzed with this method including age of onset, severity, seizure type and syndrome type.
In a study published in the September edition of Epilepsia, the researchers first applied their method to 63 families from the Epilepsy Family Study of Columbia University, an ongoing study led by Dr. Ottman that began in 1985. They examined the genetic effects on two major subtypes of epilepsy generalized epilepsy, in which seizures begin on both sides of the brain at once, and localization-related or focal epilepsy, in which the seizures begin in only one part of the brain. They found that these types clustered in families, providing evidence for significant genetic differences between these two types.
Next, in the Neurology paper, they used their approach to analyze two types of generalized seizures myoclonic and absence seizures and found evidence for distinct genetic influences on these two types. "Seizure type seems to be determined by a person's genes," Dr. Winawer says.
Then they applied their strategy to investigate the genetic influence on generalized epilepsy "syndromes," which are larger clinically defined categories that include seizure type, age of onset, severity and other information, and which have been the focus for most other investigators. Surprisingly, the researchers did not find evidence for distinct genetic influences on the different syndromes, unless the syndromes differed by seizure type. This suggests genetic differences among these syndromes may actually be driven by the seizure types that define them. Since then, the researchers have applied the same methods to a set of families from Australia and confirmed their results.
Taken together, the findings suggest that in future genetic linkage studies of generalized epilepsy which aim to find the locations of disease genes on chromosomes relative to known genetic markers a focus on seizure type in addition to syndrome type may be useful. In the meantime, the researchers are sharing their method with other researchers through a computer program Dr. Rabinowitz developed. This innovative and useful method can be applied to many different diseases. "We hope our method will help us and other researchers bring more clarity to study design by defining disease types by symptoms likely to have a genetic basis," Dr. Winawer says.