Partek will soon offer a new application that unites two commonly used genotype association approaches in order to measure genetic relatedness between individuals.
The approach is expected to improve the identification of biomarkers from genome-wide association study data because it will provide a more accurate characterization of unreported familial relationships, which would otherwise inflate the occurrence of common SNPs in the dataset and lead to misinterpretation of the data.
The method, which the St. Louis, Mo.-based bioinformatics company developed in collaboration with researchers at the Kennedy Krieger Institute, calculates how closely individuals are related by combining information about shared genetic material between unrelated individuals — or identity by state — and how this material is transmitted from parents to children — or identity by descent.
Jonathan Pevsner, the senior study author and director of bioinformatics at the Kennedy Krieger Institute, told BioInform that the approach improves upon existing genetic variation analysis tools because it provides a "graphically and visually user-intuitive" way of exploring SNP data.
The developers explain in a PLoS Genetics paper discussing the approach that it combines estimates of identity by descent between two individuals with identity-by-state plots that show genetic relationships between two unrelated people.
The combined techniques reveal "previously unknown familial relationships and population substructure in large-scale SNP data" without requiring "prior information about relatedness between individuals or haplotypes," a capability that is currently lacking in existing tools, the authors wrote.
Specifically, the developers were able to find previously unknown familial relationships and population relatedness in genomic datasets, ascertain the nature of relatedness between particular individuals, and call fewer false positives — unrelated individuals who are called as related — than existing methods.
The team expects the approach to benefit groups who are studying SNPs as part of association and linkage studies as well as those exploring genetic inheritance within families and pedigrees.
The researchers believe the method could ultimately aid the search for disease biomarkers because it will identify unknown familial relationships within large-scale genomic datasets. Such studies rely on unbiased representative sampling of populations to ensure validity. As a result, unreported familial relationships could bias the analysis because SNPs that occur between unrelated individuals are more meaningful than SNPs that occur between related individuals.
Pevsner told BioInform that researchers at Kennedy Krieger are currently using the method to find instances of inbreeding in patient populations, as well as a "variety of chromosomal aberrations" that are linked to childhood brain disorders.
This partnership builds on an existing relationship between the two groups that led to the development of two other tools — dubbed SNP duo, for pairwise comparisons of variant datasets, and SNP trio, which lets users visualize and analyze inheritance patterns between parents and children. Partek has already included these tools in its software suite, Tom Downey, president of Partek, told BioInform.
He explained that the new method extends the capabilities of the earlier tools, which identify inherited variants, to identify variants in unrelated individuals.
When the developers applied the method to publicly available datasets, "we found all variety of relatedness" in individuals who were previously believed to be unrelated as well as cases where some samples were listed under the wrong ethnicity, he said.
In the PLoS Genetics paper, the authors wrote that when they applied their method to a set of 400 supposedly unrelated individuals in a human variation panel from the Coriell Institute's Human Genetic Cell Repository, they identified "identical, parent-child, and full-sibling relationships and reconstructed pedigrees."
To ensure that the finding wasn’t a fluke, the team validated the method by applying it to data from individuals from the same family from a clinical study.
They also compared the method to the Broad Institute's PLINK, a tool for whole-genome association analysis that includes capabilities for measuring genetic relatedness, and found that it called more false positives compared to the Kennedy Krieger method.
Downey said that Partek plans to include the method in its Genomic Suite software for analyzing and visualizing microarray and next-generation sequence data.
Researchers can also download the method's source code from the Kennedy Krieger website.
Have topics you'd like to see covered in BioInform? Contact the editor at uthomas [at] genomeweb [.] com.