Scientists Develop New Algorithm that Can Better Detect Disease-Causing Genetic Variants

Scientists at Children’s Hospital of Philadelphia (CHOP) have developed a new open-source algorithm that can detect with high sensitivity disease-causing structural variants in the genome. The tool, called LinkedSV, applies the algorithm to fragments of barcoded genomic material, allowing the identification of changes in the genome that previously evaded detection.

Kai Wang, PhD, a genomics researcher at the Raymond G. Perelman Center for Cellular and Molecular Therapeutics at CHOP, developed the algorithm in collaboration with a team of scientists from CHOP’s Center for Applied Genomics, as well as the University of Pennsylvania and the Medical University Innsbruck in Austria. The method was published Dec. 6, 2019 in Nature Communications.

Current genome sequencing methods involve using so-called short-reads, fragments of genomic material that are scanned for genetic anomalies. However, those fragments are, on average, 100 to 150 DNA base pairs long. Many disease-causing structural changes within the genome are much longer than that and thus difficult to detect using short-read sequencing technologies. As a result, traditional exome sequencing methods miss approximately 50 to 70% of potential diagnoses for rare undiagnosed diseases.

In the new LinkedSV method described by Wang and his collaborators, scientists used linked-read technology, fragmenting the genome into pieces 1,000 times longer than traditional sequencing methods. The longer snippets of genomic material were dispersed into droplet partitions, and each droplet was given a barcode. DNA sequences in the droplets from the same larger fragment received the same barcode, allowing the scientists to group together sequences that reside in proximity within the genome, resulting in much longer genomic sequences. Doing so allowed the scientists to discover deletions, duplications, translocations and inversions within the genome that traditional sequencing methods had missed.

“We are suggesting that the method can help improve the diagnostic rate for undiagnosed diseases,” said Wang, noting the algorithm is already being used at CHOP to reanalyze patients whose diagnoses may have been missed previously using other sequencing techniques. “We think this is a future direction that should be pursued.”

In addition to its potential clinical use, Wang said the method is fully compatible with current sequencing procedures, requiring only the addition of the barcoding step through a microfluidic system.

“By adding a small extra cost, you may be able to analyze sequence data that is much more informative and can capture previously unseen structural changes in the human genome,” Wang said.

In addition to his CHOP position, Wang also is an Associate Professor of Pathology and Laboratory Medicine in the Perelman School of Medicine at the University of Pennsylvania.

Fang, L., Kao, C., Gonzalez, M.V. et al. “LinkedSV for detection of mosaic structural variants from linked-read exome and genome sequencing data.” Nature Communications, published online December 6, 2019. doi:10.1038/s41467-019-13397-7