Molecular diagnoses in the congenital malformations caused by ciliopathies cohort of the 100,000 Genomes Project

Abstract

Background: Primary ciliopathies represent a group of inherited disorders due to defects in the primary cilium, the 'cell's antenna'. The 100,000 Genomes Project was launched in 2012 by Genomics England (GEL), recruiting National Health Service (NHS) patients with eligible rare diseases and cancer. Sequence data were linked to Human Phenotype Ontology (HPO) terms entered by recruiting clinicians.

Methods: Eighty-three prescreened probands were recruited to the 100,000 Genomes Project suspected to have congenital malformations caused by ciliopathies in the following disease categories: Bardet-Biedl syndrome (n=45), Joubert syndrome (n=14) and 'Rare Multisystem Ciliopathy Disorders' (n=24). We implemented a bespoke variant filtering and analysis strategy to improve molecular diagnostic rates for these participants.

Results: We determined a research molecular diagnosis for n=43/83 (51.8%) probands. This is 19.3% higher than previously reported by GEL (n=27/83 (32.5%)). A high proportion of diagnoses are due to variants in non-ciliopathy disease genes (n=19/43, 44.2%) which may reflect difficulties in clinical recognition of ciliopathies. n=11/83 probands (13.3%) had at least one causative variant outside the tiers 1 and 2 variant prioritisation categories (GEL's automated triaging procedure), which would not be reviewed in standard 100,000 Genomes Project diagnostic strategies. These include four structural variants and three predicted to cause non-canonical splicing defects. Two unrelated participants have biallelic likely pathogenic variants in LRRC45, a putative novel ciliopathy disease gene.

Conclusion: These data illustrate the power of linking large-scale genome sequence to phenotype information. They demonstrate the value of research collaborations in order to maximise interpretation of genomic data.

Keywords: and neonatal diseases and abnormalities; congenital; diagnosis; genetics; genomics; hereditary; medical.

Figure 1

Research analysis workflow that (A) describes steps to filter genomic data, (B) analyse putative pathogenic variants and (C) classify variants then assign diagnostic confidence. ACMG, Association for Clinical Genomic Science; DDG2P, Development Disorder Genotype - Phenotype Database; GEL, Genomics England; IGV, Integrative Genomics Browser; RMCD, rare multisystem ciliopathy disorders; SNV, single nucleotide variant; SV, structural variant; VEP, variant effect predictor; VUS, variant of uncertain significance.

Figure 2

Comparison of diagnostic reporting outcomes between gel GMC exit reports (A) and research diagnostic outcomes (B) for the 83 probands in the CMC cohort. (C) Research molecular diagnoses according to recruitment category. Genes with identified potentially causative variants are grouped according to whether they are known to be associated with ciliopathies or not. A ‘+’ is used where participants had potentially causative variants in more than one gene contributing to their clinical features (additional gene(s) are included in brackets). Diagnostic confidence for each research molecular diagnosis is shown in table 1. Detailed variant information, including whether the gene variants(s) are thought to be a full or partial match to phenotype, is provided in online supplemental table 4. BBS, Bardet-Biedl syndrome; CMC, congenital malformations caused by ciliopathies; GEL, Genomics England; GMC, Genomic Medicine Centre; JBTS, Joubert syndrome; RMCD, rare multisystem ciliopathy disoder.

Figure 3

IGV captures of structural variants identified among participants of the congenital malformations caused by ciliopathies cohort. First, an untiered ALMS1 SV identified in participant #29 was initially called a tier 2 ALMS1 missense variant. Closer inspection on IGV determined that this was an indel (92 nucleotide deletion and 31 nucleotide insertion) leading to a splice acceptor change at the beginning of exon 6 (A). Our filtering pipeline identified a second untiered ALMS1 frameshift variant, completing the molecular diagnosis of Alström syndrome. Three larger heterozygous deletions were identified through manual IGV inspection of whole gene loci when searching for second hits in probands with potentially causative SNVs. An untiered 13.3 kb deletion in PIBF1 (also known as CEP90) (B) was identified in a proband with an untiered novel missense variant (proband #42). An untiered 4.5 kb deletion in BBS1 (C) was found in a proband with an untiered, ClinVar pathogenic missense variant (proband #69). Finally, a 2.7 kb deletion in CSPP1 (D) was found in a proband with a predicted splice donor loss (SpliceAI DS_DL 0.79) (proband #41). This CSPP1 deletion was only seen in ~30% of reads in the proband but in ~50% of reads in their father. SNV, single nucleotide variant.