CHARLOTTE, NC (GenomeWeb News) – Researchers involved with the National Institutes of Health Undiagnosed Diseases Program are using an exome sequence analysis approach that they call "extreme novel filtering" to find disease-associated mutations behind conditions represented by just one or a few patients.
By comparing each patient's exome sequence data with the exomes of his or her immediate family members — along with sequence databases and exomes from UDP patients with other conditions — it is possible to narrow in on candidate mutations that could not be found looking at a single patient exome, explained Cornelius Boerkoel.
The UDP researchers are taking this a step further by focusing only on genetic alterations found in parts of the exome with good coverage, he added, which eliminates stretches of sequence where a significant proportion of spurious variants tend to crop up.
Boerkoel, director of the UDP translational lab at the National Human Genome Researcher Institute, provided an update on the UDP and the strategies it's using at the American College of Medical Genetics and Genomics annual meeting here today.
The Undiagnosed Disease Program was launched in 2008 with the goal of finding the genetic causes behind conditions that confound conventional diagnostic tests. In this January's issue of Genetics in Medicine, the team outlined its success rate in diagnosing disease over the first two years of the program.
Cases selected for inclusion in the program undergo extensive phenotypic and clinical evaluations to ensure that each individual's symptoms and genetic features are distinct from those present in known diseases, Boerkoel explained.
Once UDP researchers have evidence that a disease has genetic roots but cannot be diagnosed by existing methods, they go on to do their own genetic analyses using microarray and sequencing-based approaches.
While SNP analyses have been sufficient to diagnose some of the diseases the team has tackled so far, Boerkoel said that much of the genetic screening that's currently underway is based on exome sequencing, or, in some cases, whole genome sequencing.
That has made it necessary to come up with effective filtering strategies to work back from the 20,000 or so variants found in each person's exome to get to alterations present only in individuals with a given disease.
To make this feasible, researchers have been doing exome sequencing not only on patients, but also parent and affected or unaffected siblings, Boerkoel explained. Along with comparisons between patient exomes and exome sequences from family members, they have been able to find novel variants by filtering UDP patient exomes against one another, since each condition is so rare.
They also have been weeding out variants present in the ClinSeq and 1000 Genomes Project databases as well as variants found outside of exome regions with good coverage — a step that has helped in removing erroneous variants that result from poor alignment in certain exome regions.
When they applied this extreme novel filtering approach to 22 families containing 27 patients with undiagnosed diseases, for example, the researchers found 84 candidate mutations. They then pared this set of candidate genetic changes down even further with Sanger sequencing-based validation studies testing for the presence or absence of the possible mutations in unaffected parents or siblings.
Overall, Boerkoel said, the approach made it possible to find de novo candidate mutations in the 22 families with a success rate of more than 50 percent.
The team is putting together a pipeline for analyzing around 150 more exomes from families enrolled in the UDP. Candidate mutations detected in patient exomes are ultimately tested experimentally to determine whether they are authentic disease-causing alterations.