USTAR Center for Genetic Discovery Improves Disease Identification Tool at VAAST Hackathon
A child is born in a hospital and has a life-threatening illness that appears to be genetic in nature, but none of the attending physicians can figure out what it is. With a new software tool called the Variant Annotation, Analysis and Search Tool (VAAST), arriving at a diagnosis may become as simple as performing a Google search.
The Utah Science Technology and Research initiative’s (USTAR) Genetic Discovery team at the University of Utah (U of U) hosted a weeklong hackathon for VAAST developers to expand the features and usability of the VAAST tool. The hackers spent the week hashing out improvements at all stages of the VAAST pipeline that starts at the clinic with a sick patient and ends with a diagnosis treatment plan.
The development group was comprised of Genetic Discovery co-director Mark Yandell’s research group at the U of U, former U of U graduate students and post-doctoral researchers, representatives from San Francisco-based Omicia, a personalized medicine company focused on the interpretation of genome sequences for clinical applications, and researchers from Houston-based MD Anderson, a leading cancer-care hospital.
VAAST is a probabilistic search tool that identifies damaged genes and their disease-causing variants in personal genome sequences. It is both an academic tool, licensed through the U of U, and a commercial tool that’s imbedded within Omicia through a product called Opal, a clinical decision support tool.
Yandell said by analyzing a genome, the VAAST tool anticipates which gene might be broken. VAAST has already identified the genetic basis of such diseases in Utah, and will be deployed routinely in the health care system within the next five to 10 years.
“As the tool has grown and we’ve seen it in operation on real data, we’ve not only discovered things like new disease genes and new genetic diseases, but we’ve also gained a greater understanding of human genomes and how they evolve and how they break,” said Yandell. “We’re always trying to incorporate those new basic findings into the tool, so that’s another big part of what we’re doing here right now.”
Barry Moore, research scientist for the genetic discovery team, said each time the group has published on VAAST, they have done an extensive series of testing as part of the publication effort.
“Those tests typically focus on the accuracy of the underlying algorithm,” said Moore. “We’ll generate data sets by spiking known disease alleles into existing publically available exomes or genomes, and we’ll see at what rate we recover those.”
Recently, VAAST was used to discover Ogden Syndrome, a disorder found in infant boys that produces an aged appearance, stunted growth and heart arrhythmias. Over the course of a few generations in one family, five boys died before their first birthday. Follow up studies with the family and the families medical history revealed that they had been bringing in a series of sick children over several generations in Utah hospitals. The children were born relatively healthy and then died within the first few months of life.
“VAAST discovered the gene involved,” said Yandell. “It didn’t just discover the cause of this genetic disease, but simultaneously discovered a new genetic disease and the cause of it. To date the disease is untreatable, but just the knowledge of who in the family carries the allele and who doesn’t is crucially important, for peace of mind if nothing else.”
A major goal for the hackers at this week’s developer’s meeting was to improve the power of the VAAST tool for cancer research. The group discussed the development of software tools that can analyze an individual’s cells before and after they become cancerous to see what mutations are making the cells invasive and ultimately predict how the cancer cells will respond to chemotherapy and other treatments.
Another focus of the group was to develop a new type of tool for detecting infectious diseases, a project done in collaboration with the Department of Pathology, ARUP laboratory, and the Center of Disease Control (CDC) to analyze data collected as part of atypical childhood pneumonia.
“If someone is sick with an intestinal bug, flu bug or anything else, we can sequence body fluids (snot and vomit) to try and find out what’s in there,” said Yandell. “For example, as part of a study called EPIC, the CDC collected specimens from thousands of severely ill children under two years old admitted to emergency rooms. Many tested negative for all known causes of pneumonia. The idea is to sequence body fluids to see if we can figure out what might be making them ill, like a virus or bacteria.”
Moore says the group is also focused on expanding the functionality of the VAAST tool beyond human disease.
“We always tried to write VAAST so it was agnostic to what kind of data set you run it on,” said Moore. “There are certainly advantages to the nonhuman model organism, stuff from a scientific point of view that you can experiment on. There’s a great scientific and economic opportunity there.”
In collaboration with Mike Shapiro, professor of biology at the U of U, Yandell said VAAST has recently been extended to pigeons to discover genetic causes for traits that Charles Darwin used to illustrate the Origin of Species. They teased apart an interacting network of genes that control color in domestic pigeons.