Unraveling undiagnosed rare disease cases by HiFi long-read genome sequencing

Abstract

Solve-RD is a pan-European rare disease (RD) research program that aims to identify disease-causing genetic variants in previously undiagnosed RD families. We utilized 10-fold coverage HiFi long-read sequencing (LRS) for detecting causative structural variants (SVs), single-nucleotide variants (SNVs), insertion-deletions (indels), and short tandem repeat (STR) expansions in previously studied RD families without a clear molecular diagnosis. Our cohort includes 293 individuals from 114 genetically undiagnosed RD families selected by European Reference Network (ERN) experts. Of these, 21 families were affected by so-called "unsolvable" syndromes for which genetic causes remain unknown and for which prior testing was not a prerequisite. The remaining 93 families had at least one individual affected by a rare neurological, neuromuscular, or epilepsy disorder without a genetic diagnosis despite extensive prior testing. Clinical interpretation and orthogonal validation of variants in known disease genes yielded 12 novel genetic diagnoses due to de novo and rare inherited SNVs, indels, SVs, and STR expansions. In an additional five families, we identified a candidate disease-causing variant, including an MCF2/FGF13 fusion and a PSMA3 deletion. However, no common genetic cause was identified in any of the "unsolvable" syndromes. Taken together, we found (likely) disease-causing genetic variants in 11.8% of previously unsolved families and additional candidate disease-causing SVs in another 5.4% of these families. In conclusion, our results demonstrate the potential added value of HiFi long-read genome sequencing in undiagnosed rare diseases.

Figure 1.

HiFi LRS in a unique cohort of 293 individuals from 114 RD families. The study cohort consists of two subcohorts: the “unsolvables” (families affected by clinically well-recognizable syndromes for which the cause is yet unknown) and the “unsolved” (families affected by a rare neurological, neuromuscular, or epilepsy disease). All patients were recruited via four European Reference Networks and subsequently sequenced using a single SMRT cell of sequencing data per individual. Genome-wide calling of SVs and SNVs was conducted, and STRs were genotyped at 56 known disease-associated loci. (ERN) European Reference Network, (BND) breakend call, (INH) inherited variant, (DNM) de novo mutation.

Figure 2.

Visualization of the TUBA1A de novo missense variant in P0185637 using IGV, and a pedigree of the family. The variant had earlier been described as a cause of lissencephaly. Healthy family members do not carry the variant. Sequenced individuals are marked with an asterisk (*) in the pedigrees.

Figure 3.

Visualization of disease-causing SVs in the “unsolved” subcohort in the form of cartoons and/or IGV screenshots, along with corresponding pedigrees. In two unrelated male patients (P0078963 in A,B, P0695060 in C,D) with muscular dystrophy, we found X-Chromosomal inversions (A–D). In both cases, DMD is disrupted (A,C), in one a second gene disruption adds to the phenotype (C). In a father and son with hereditary spastic paraplegia, we detected a deletion of REEP1 exon 6 (E–H). The deletion in P001782 and P0011781 is shown here as a cartoon (E) and as a screenshot in IGV (F). The deletion was also visualized by agarose gel electrophoresis, which confirms that both patients are heterozygous for the deletion (G). The pedigree of the family is shown in H. In a patient with adult-onset distal myopathy, a 65 kb duplication involving MYOT (I) was confirmed to be in tandem by LRS (J). The pedigree of the family is shown in K. Sequenced individuals are marked with an asterisk (*) in the pedigrees (B,D,G,H,K). (MD) Muscular dystrophy, (AD-HSP) autosomal dominant hereditary spastic paraplegia.

Figure 4.

Visualizations were produced using the PacBio TRGT tool and pedigrees for the families with pathogenic STR expansions. In siblings P0016368 and P0018504, a heterozygous GGCCTG expansion in NOP56 was detected (A). In another family, an expansion including the motifs GGCCTG and CGCCTG in NOP56 was detected in one generation (P0018996), and the STR expansion was subsequently also identified in the mother (B). In patient P0016356 and their father, we identified heterozygous STR expansion DAB1, including both ATTTT and ATTTC motifs (C). In another patient, we identified homozygous STR expansions in RFC1 (D). Alleles are denoted by “A1” and “A2.” Sequenced individuals are marked with an asterisk (*) in the pedigrees. (AD-ATX) Autosomal dominant ataxia.

Figure 5.

Visualization of disease-causing SNVs and indels in the “unsolved” subcohort in the form of IGV screenshots, along with corresponding pedigrees. In a sporadic patient with suspected titinopathy, we identified a deep-intronic variant in DMD (A). The nonaffected sibling did not carry the variant (A,B). In a duo consisting of an affected mother and affected daughter with HSP, we identified a noncanonical splice site variant in SPAST (C,D). In a patient with titinopathy (P063122), a maternally inherited and a de novo variant had been identified earlier (C–F). The two variants are located 109 kb apart, but the alleles were successfully phased through the entire region by LRS (G). The reads are colored by haplotag; pink and light blue, or yellow and purple represent different alleles in A, C, E, and G. Unphased reads, such as X-Chromosomal reads in males, are shown in gray (A,E). Sequenced individuals are marked with an asterisk (*) in the pedigrees (B,D,F). (BMD) Becker muscular dystrophy, (AD-HSP) autosomal dominant hereditary spastic paraplegia.

Figure 6.

Visualization of candidate disease-causing SVs in the “unsolved” cohort. In a sporadic patient (P0093700) with arthrogryposis multiplex congenita, we detected an X-Chromosomal tandem duplication (A–D). The duplication spans from intron 1 of MCF2 to intron 2 of FGF13 and also includes F9 (A). The result of the duplication is a hypothetical fusion gene, including FGF13 exons 1–2 and MCF2 exons 2–29 (B). The duplication was validated by PCR and agarose gel electrophoresis (C) using primers targeting the breakpoints of the duplication, and a combination of the MCF2 forward and FGF13 reverse primers (primer pair 1, A,C). In a sporadic patient with psychomotor development delay (P0021581), we detected a deletion of PSMA3 exons 9–11, here shown as a cartoon (E) and as a screenshot using IGV (F). The deletion was validated by agarose gel electrophoresis (G) and Sanger sequencing. The index P0021581 has three healthy siblings who do not carry the deletion (H). In a sporadic female patient (P0537031) with congenital malformation syndrome, we detected a 5 Mb tandem duplication on Chromosome 4, visualized here as a cartoon (I). The pedigree is shown in J. In a sporadic male patient (P0016165) with autosomal dominant spastic paraplegia, with a similarly affected father, we detected a 300 kb tandem duplication on Chromosome 2, visualized here as a cartoon (K). The pedigree is shown in L. In a family with an affected mother and son (P0847234 and P0958540), we detected an intronic variant in REEP1 (M). The reads are colored by haplotag; pink and light blue represent different alleles in M. The index patient also has an affected uncle, whose sample was not sequenced (N). Sequenced individuals are marked with an asterisk (*) in the pedigrees (D,H,J,L,N). (NMD) Neuromuscular disorder, (CMS) congenital malformation syndrome, (AD-HSP) autosomal dominant hereditary spastic paraplegia.