UCSD Study conducted more sense genome data

UCSD Study conducted more sense genome data:

LA JOLLA — The ever-faster methods of genome sequencing have caused a bit of data gridlock among researchers looking into the links between DNA sequences and disease: There's too much information to easily make sense of.

A team of UC San Diego researchers says it has found a way to better interpret the mountains of genomic data to track the associations. Their study was published April 25 in two papers in the journal PLOS Genetics.

The researchers have developed a statistical method to toss out the vast amounts of data likely to be irrelevant. Scientists can then focus more closely on the information likely to be related to diseases. By better understanding the disease process at its genetic roots, researchers hope to speed development of better therapies.

The study, led by UCSD professor Anders Dale, concerns one-letter variants in the DNA alphabet. These variants are called single-nucleotide polymorphisms, or SNPs. They can occur inside genes, or in the much larger stretch of DNA found outside genes. They can be found by scanning the entire genome to find variants associated with disease; these are called genome-wide association studies, or GWAS. The human genome contains about 3 billion letters, so a change of one letter is comparatively tiny.

Modern genomic technology has identified thousands of these SNPs. This discovery has turned on its head the common notion that a certain mutation occurs "for" a certain disease. Instead of just one mutation in a disease, there are most often many mutations.

Sometimes, people will have disease-associated SNPs but not develop the disease. That fact makes the genetic tests being sold commercially of questionable value in disease prediction. Also, genetic variations may be associated with more than one disease, a condition known as pleiotropy.

Add this all up and it means there's no obvious way to associate most of these variants with diseases in any reliable fashion. That's where the UCSD-led team says its method can make a difference.

"It's increasingly evident that highly heritable diseases and traits are influenced by a large number of genetic variants in different parts of the genome, each with small effects," said Dale, a professor in the departments of Radiology, Neurosciences and Psychiatry at the UCSD School of Medicine, in a statement on the studies. "Unfortunately, it's also increasingly evident that existing statistical methods, like genome-wide association studies that look for associations between SNPs and diseases, are severely underpowered and can't adequately incorporate all of this new, exciting and exceedingly rich data."

The study developed a shortcut to finding patterns of these SNPs that bunch up in certain diseases. The conclusion is made plain in the title of one of the papers, which stated in part, "All SNPs are not created equal."

"We hypothesize that all SNPs in a GWAS are not exchangeable, but come from pre-determinable categories with different distributions of effects," the study stated. "This implies that some categories of SNPs are enriched, i.e. are more likely to be associated with a phenotype than others."

The method also includes pleiotropic information, recognizing that diseases may share common genetic causes.