Skip to main content
Premium Trial:

Request an Annual Quote

Solve-RD's Genomic Reanalysis Yields New Diagnoses for Hundreds of Rare Disease Patients

Premium

NEW YORK – Systematic reanalysis of existing genomic and phenotypic data from more than 6,000 unsolved rare disease patient cases in Europe has led to hundreds of new genetic diagnoses, according to a study published in Nature Medicine on Friday.

Led by researchers from the pan-European Solve-Rare Diseases Consortium (Solve-RD), the landmark study highlights the utility of reanalyzing patients' genomic data and the importance of interdisciplinary collaboration for ending diagnostic odysseys.

"The most important lesson is, there is huge intrinsic value in data that was produced yesterday, and it may be worth looking at it again," said Alexander Hoischen, a genomics researcher at Radboud University Medical Center in the Netherlands and a corresponding author of the study.

Kicked off in 2018 with €15.4 million ($16.7 million) in funding from the European Union, Solve-RD is a large-scale, ​​pan-European effort that aims to boost the diagnostic rate for individuals affected by a rare disease. As part of the project, over 300 experts — including clinicians, laboratory geneticists, and translational researchers — from dozens of research groups joined forces to tackle those hard-to-solve cases.

The publication covers 6,447 individuals with previously undiagnosed rare diseases from 6,004 families. Of these, 3,592, or 56 percent, were male while 2,855, or 44 percent, were female. Additionally, the analysis included data from 3,197 unaffected family members.

The reanalysis focused on four categories of rare diseases, each spearheaded by a corresponding European Reference Network (ERN). Among the families, 2,271 had rare neurological diseases; 1,857 had malformation syndromes, intellectual disability, and other neurodevelopmental disorders; 1,517 had rare neuromuscular diseases; and 359 had suspected hereditary gastric and bowel cancers.

The project considered more than 9,870 genomic datasets from these families following quality control, including around 520 genomes and over 9,350 exomes, all previously generated and analyzed through local diagnostic or research efforts. Along with the sequencing data, phenotypic descriptions and pedigree information for the participants were included in the study.

According to Steven Laurie, a bioinformatician at the National Center for Genomic Analysis (CNAG) in Spain and a coauthor of the study, all sequencing data included in the study were short reads, with the majority generated using an Illumina platform and some produced on the DNBSeq sequencer from MGI Tech.

To tackle these cases, the researchers organized themselves into a two-level expert review framework within each ERN, including a data analysis task force and a data interpretation task force, to leverage bioinformatics and clinical genetics expertise, respectively.

Within the data analysis task force, the researchers further divided themselves into different working groups to home in on certain types of variants. Overall, the study reanalyzed a broad range of genomic alterations, including single nucleotide variants, short insertions and deletions, non-canonical splice variants, homoplasmic and heteroplasmic mtDNA variants, copy number variants, structural variants, mobile element insertions, and short tandem repeat expansions.

Meanwhile, the data interpretation task force within each ERN generated a curated list of genes implicated in diseases studied by their group. "While the bioinformaticians were very busy with curating the data, on our side, we generated gene lists depending on the disease and the phenotype," said Richarda de Voer, a cancer genomics researcher at Radboud UMC and a coauthor of the study.

These gene lists were then used as a reference to help identify potentially pathogenic variants within the patients, de Voer noted.

Overall, the Solve-RD researchers obtained genetic diagnoses for 506 families, translating to a diagnostic yield of more than 8 percent.

Of the 552 disease-causing variants identified, 464, or 84 percent, were single nucleotide variants (SNVs). Of these, 67 were located in recently published novel disease genes, 187 were variants recently reclassified in ClinVar, and 210 were reclassified by consensus expert decision within Solve-RD. The remaining solved cases were caused by variants that were non-SNVs.

In parallel to Solve-RD's efforts, many local research groups that contributed cases to the initiative also continued to analyze their cases of interest. Such analysis helped 249 additional families reach a diagnosis. Cases that were solved through ad hoc expert review were reported to Solve-RD and not interpreted further as part of the systematic reanalysis effort.

As part of the study, the Solve-RD researchers further investigated the clinical actionability of the results by considering medications or interventions included in three databases: IEMbase, Treatabolome, ClinGen, and international cancer guidelines. They identified variants in potentially actionable genes for 73 individuals, or more than 14 percent of those diagnosed.

At least 16 families, including those with neurological diseases, neuromuscular disorders, and genetic tumor risk syndromes, have received relevant treatments as a result, according to the study.

The paper represents data from the first two phases of Solve-RD, according to Hoischen. Since then, the consortium has investigated thousands more cases, more than 20,000 datasets from over 13,000 families in total. The analysis for the third phase is still ongoing, and results will be published as they become available, he noted.

For some of those patients who did not receive a diagnosis through the reanalysis effort, the researchers are working to perform additional experiments, such as RNA sequencing and whole-genome sequencing, to reach an answer.

Besides reanalysis, another important component of Solve-RD is to explore the utility of new omics tools and technologies for solving rare diseases. For instance, some consortium collaborators, including Hoischen's group, recently published a study in the American Journal of Human Genetics showcasing their findings using Pacific Biosciences HiFi long-read genomes to help identify difficult-to-detect, clinically relevant variants.

All the phenotypic information and genetic variants analyzed in the Solve-RD study are accessible to registered users using the RD-Connect Genome-Phenome Analysis Platform (GPAP). Additionally, all raw and processed data files are available from the European Genome-Phenome Archive. By making all the data available, Hoischen said, Solve-RD aims to build a valuable resource for rare disease researchers.

While funding for Solve-RD has ended, the infrastructure and collaborative framework developed by the project may serve as a blueprint for future large-scale rare disease endeavors, such as the European Rare Diseases Research Alliance (ERDERA) launched last September, Hoischen said. Coordinated by France’s INSERM in partnership with over 180 research groups — including key scientists from Solve-RD — the program, with an estimated budget of €380 million, may also offer new opportunities for some Solve-RD patients who are still waiting for an answer.

“The constant ping-pong between disease experts and genomics and technical experts is really what was showcased in this manuscript, and in the Solve-RD effort in general,” Hoischen said. “The joint effort and collective brain power really makes the difference.”