Summary
A bioRxiv preprint presents GAVURD, a trio whole-genome analysis system for finding potentially pathogenic variants outside protein-coding regions. In a proof-of-concept involving 10 unresolved rare-disease cases, it prioritised six potentially causal variants for further study.
Researchers at the Children's Hospital of Philadelphia have presented GAVURD, a computational system designed to prioritise potentially pathogenic variants in the non-coding genome of people whose rare diseases remain genetically unresolved.
In a proof-of-concept analysis, the system examined trio whole-genome sequencing data from 10 unresolved rare-disease probands and identified six non-coding variants as potentially causal. The authors describe these variants as candidates for additional functional follow-up rather than established molecular diagnoses. The work is available as a bioRxiv preprint.
Why non-coding variants matter
Rare-disease analysis often focuses on the protein-coding parts of genes, because changes in these regions can directly alter the proteins cells make. Exome sequencing examines mainly these coding regions, while whole-genome sequencing also detects DNA between and around genes. That broader sequence contains regulatory elements that can influence when, where and how strongly genes are active.
A non-coding variant can therefore affect disease biology without changing a protein’s sequence. The difficulty is linking a variant in a regulatory region to the gene or genes it influences. Physical distance along the DNA sequence is not always a reliable guide to these relationships because the genome folds inside the cell.
GAVURD uses topologically associated domains, or TADs, to help make those links. A TAD is a three-dimensional region of the genome in which DNA segments interact more frequently with one another. Incorporating this organisation allows the system to connect candidate non-coding variants with human disease genes that occupy the same genomic neighbourhood.
How GAVURD ranks candidates
The system uses alignment data from trio whole-genome sequencing. In this context, a trio generally consists of an affected child and both parents, allowing researchers to distinguish variants inherited from a parent from de novo variants that arise in the child.
According to the preprint, GAVURD applies approaches for identifying rare inherited and de novo variants, links those variants to disease genes using TAD information, and then ranks them according to overlap between the patient’s observed features and the known features associated with candidate genes.
This creates a shorter, more informed list for researchers to investigate. Instead of treating every rare non-coding change as equally important, the system combines inheritance, genomic organisation and phenotype-related evidence to prioritise variants with several signals supporting pathogenicity.
A proof of concept, not a completed diagnosis
When the researchers applied GAVURD to 10 probands with unresolved rare disease, the analysis implicated six potentially causal non-coding variants. The result shows how whole-genome data can be re-examined beyond the protein-coding regions that commonly receive the most attention in clinical interpretation.
The study is a proof of concept based on a small group of cases, and it is reported in a preprint rather than a peer-reviewed publication. The six variants are presented as putative candidates whose biological effects require further functional investigation. Such studies would examine whether a variant changes gene regulation in a way that is consistent with the patient’s condition.
The immediate value of GAVURD is therefore prioritisation: it can identify high-value candidates for laboratory and clinical-genomics teams to test. Its diagnostic performance across larger and more diverse unresolved rare-disease cohorts remains to be established.