Short-read whole genome sequencing identifies causative variants in most individuals with previously unexplained aniridia

Background Classic aniridia is a highly penetrant autosomal dominant disorder characterised by congenital absence of the iris, foveal hypoplasia, optic disc anomalies and progressive opacification of the cornea. >90% of cases of classic aniridia are caused by heterozygous, loss-of-function variants affecting the PAX6 locus. Methods Short-read whole genome sequencing was performed on 51 (39 affected) individuals from 37 different families who had screened negative for mutations in the PAX6 coding region. Results Likely causative mutations were identified in 22 out of 37 (59%) families. In 19 out of 22 families, the causative genomic changes have an interpretable deleterious impact on the PAX6 locus. Of these 19 families, 1 has a novel heterozygous PAX6 frameshift variant missed on previous screens, 4 have single nucleotide variants (SNVs) (one novel) affecting essential splice sites of PAX6 5′ non-coding exons and 2 have deep intronic SNV (one novel) resulting in gain of a donor splice site. In 12 out of 19, the causative variants are large-scale structural variants; 5 have partial or whole gene deletions of PAX6, 3 have deletions encompassing critical PAX6 cis-regulatory elements, 2 have balanced inversions with disruptive breakpoints within the PAX6 locus and 2 have complex rearrangements disrupting PAX6. The remaining 3 of 22 families have deletions encompassing FOXC1 (a known cause of atypical aniridia). Seven of the causative variants occurred de novo and one cosegregated with familial aniridia. We were unable to establish inheritance status in the remaining probands. No plausibly causative SNVs were identified in PAX6 cis-regulatory elements. Conclusion Whole genome sequencing proves to be an effective diagnostic test in most individuals with previously unexplained aniridia.


INTRODUCTION
Historically the molecular genetic investigation of Mendelian disorders has focused on sequencing of the coding regions of causative genes often in combination with genomic copy number analysis.Depending on the phenotypic specificity of the disease under investigation, these tests could involve sequencing of a single gene, a panel of genes or whole exome analysis.One of the motivations for restricting diagnostic analysis to the coding regions of genes has been the availability of wellcharacterised and validated approaches to predict the consequence of each variant and to assign a confidence to its pathogenicity. 1 2 The wider adoption of whole genome sequencing (WGS) as a diagnostic tool, 3 together with the guidelines that aim to standardise the interpretation of variants outside the coding regions of genes, provides an opportunity to increase the utility of diagnostic genetic analyses. 4Here we have used short-read WGS to try to identify causative variants in individuals with aniridia in whom no diagnosis was found by prior molecular genetic testing approaches.Classic aniridia has major advantages in such a study as a Mendelian disease in which the phenotype in early childhood (congenital absence of the iris with foveal hypoplasia) has a high (~0.9)positive predictive value for detecting a heterozygous loss-of-function mutation at a single locus (PAX6). 5

WHAT IS ALREADY KNOWN ON THIS TOPIC
⇒ Aniridia is a strikingly specific phenotype, carrying a 90%-95% positive predictive value for PAX6 haploinsufficiency.

Developmental defects
PAX6 encodes a dosage critical transcription factor that is essential for vertebrate eye and brain development, 6 and many cis-regulatory elements (CRE) controlling its expression during development have been functionally characterised. 7 8Diagnostic analysis of this locus has been in routine clinical use now for 30 years, providing a very large data set of causative variants and established disease mechanisms. 5s this study will show, WGS is able, with reasonable sensitivity, to identify causative variants in individuals with a clinical diagnosis of aniridia who have previously, often repeatedly, tested negative for mutations in the coding region of PAX6.

Clinical research participants
This project used clinical information and biological samples from individuals referred to the Medical Research Council (MRC) Human Genetics Unit Eye Malformations Study.All affected individuals had classical aniridia; those with significant additional ocular phenotypes were excluded, such as severe microphthalmia or severe congenital corneal opacification.Pseudonymised research participant identifiers (RPIDs), relevant clinical features and molecular analyses performed prior to this study are provided in table 1.

Preparation of genomic DNA and quality control
The quality and concentration of patient genomic DNA (gDNA) samples were assessed by agarose gel electrophoresis, NanoDrop 1000 spectrophotometry (Thermo Fisher Scientific, Inchinnan, UK), and/orQubit 3 fluorometer high sensitivity (HS) assay (Invitrogen, Thermo Fisher Scientific).In the case of one family trio, the DNA for the parents was extracted from patient-derived lymphoblastoid cell lines (LCLs).All probands had had prior Sanger sequencing of the coding regions of PAX6 and MAB21L1.

Whole genome sequencing
WGS was performed by BGI (New Territories, Hong Kong) for 9 samples and Edinburgh Genomics (Edinburgh, UK) for the remaining 42 samples.
Detailed methods for massively parallel sequencing library preparation are found in online supplemental information.In brief, gDNA was sheared using a Covaris ultrasonicator, and the fragments were A-tailed, size selected and adaptor ligated prior to PCR amplification.Libraries were clustered onto a flow cell for sequencing using HiSeqX (Illumina).

Structural variant calling IGV visualisation
Direct inspection of the aligned WGS data was performed to detect structural variant (SV) at known aniridia loci, using IGV (Integrative Genomics Viewer, Broad Institute, Massachusetts, USA), 13 via visualisation of breakpoints and coverage.The gene and regulatory regions of PAX6 were examined in all cases.Where PAX6 was negative, FOXC1 and PITX2 were included with their regulatory regions, and MAB21L1; in some cases, this was expanded to additional loci (FOXE3, RARB, ADAMTSL1, CYP1B1); for a list of coordinates, see online supplemental table S1.In IGV, reads were coloured both by insert size (to detect breakpoints of deletions/insertions) and by pair orientation (to detect breakpoints of chromosomal rearrangements such as inversions).

Bioinformatic SV calling
CNVs were called for each family with Canvas V.1.38 14using the 'SmallPedigree-WGS' workflow. 15SVs were called with Manta V.1.3.2. 16 17CNV/SV overlapping all genes in the EyeG2P data set (described in the Variant filtering section) were examined.

De novo analysis
For parent-child trios, short variants arising de novo in the child were identified using VASE 18 with the following criteria: read depth ≥10 in parents and the child, genotype quality score ≥30 in the child and ≥20 in both parents, variant allele frequency ≥0.3 in the child and <0.05 in both parents, with a maximum of one variant allele called at site and variant allele absent from gnomAD V.3.0.Potential de novo variants were subsequently filtered to exclude low complexity, telomeric and centromeric regions, and to select only variants within coding regions, splice regions (exonic positions within 3 bp of an intron/exon junction or intronic positions within 8 bp of an intron/exon junction) or intronic variants with a SpliceAI 19 delta score ≥0.5.

Variant filtering
From all the variants identified in an individual, we selected only those that are rare, predicted to be functional and potentially relevant to eye disorders by using the G2P plugin 20 21 in VEP (V.90.1 (16)) and the Eye Gene Panel (https://www.ebi.ac.uk/gene2phenotype/downloads; accessed 29 August 2018).In short, we extracted only variants satisfying the inheritance requirements of the genes in the Eye Gene Panel, with minor allele frequency (MAF) in public databases <0.0001 for monoallelic and X-linked genes and MAF <0.005 for biallelic genes and annotated by VEP to have one of the following consequences: stop gained, stop lost, start lost, frameshift variant, inframe insertion/deletion, missense variant, coding sequence variant, initiator codon variant, transcript ablation, transcript amplification, protein altering variant, splice donor/acceptor variant (ie, canonical splice site) or splice region variant (ie, either within 1-3 bases of the exon or 3-8 bases of the intron).

Cis-regulatory variant analysis
The GRCh38 coordinates of 35 different CREs chosen for analysis are given in online supplemental table S4.A BED file was created from this table and BEDTools 22 was then used to extract the PAX6 CRE variants from the cohort VCF file.Subsequent filtering used the gnomAD 23

Experimental analysis of splice variants
RT-PCR and nested PCR of LCL-derived RNA Patient-derived LCLs were recovered from liquid nitrogen storage.The cells were grown in suspension in Roswell Park Memorial Institute (RPMI) media containing 15% fetal calf serum (FCS) and penicillin/streptomycin, and incubated at 37°C/5% CO 2 .RNA was extracted from LCLs following the principles of the phenol-chloroform method 25 using TRIzol reagent (Invitrogen) following the manufacturer's instructions.

Developmental defects
DNase treatment was performed using TURBO DNase kit (Invitrogen).cDNA was obtained from total RNA using the Super-Script First-Strand Synthesis System for RT-PCR kit (Invitrogen).
The products were run on a 2% low melting point agarose gel with ethidium bromide.Bands of interest were excised and a gel DNA extraction was performed (Zymoclean Gel DNA Recovery Kit, Zymo Research, Freiburg im Breisgau) and sent for Sanger sequencing to look for mis-splicing.

Assembling the cohort
DNA samples were available for 443 individuals with aniridia from 347 families recruited to the MRC Human Genetics Unit Eye Malformation Study.Forty-five families considered to be 'PAX6-negative' on previous screening were considered for inclusion in the WGS analysis.Of the 45 families, 3 were excluded on the basis of quality or quantity of the stored DNA and 4 were excluded following identification of various gene-disruptive variants via amplicon-based resequencing of all PAX6 coding exons in all probands.Another family, RPID 1201, was excluded when a deletion of a critical cis-regulatory region 3′ of the PAX6 gene (online supplemental figure S1) was found using droplet digital PCR across the PAX6 locus following prescreening of 13 randomly chosen unrelated probands for CNVs.DNA samples from a final cohort of 39 affected individuals from 37 families together with 12 unaffected relatives were sent for WGS.The family structures comprised 29 singleton probands, 6 trios (proband plus both unaffected parents) and 2 affected relative pairs (online supplemental figure S2).The proband phenotypes and molecular analyses of PAX6 locus performed prior to this study (including the results from the referring centre) are detailed in table 1.
The following analyses of WGS data to identify sequence variants and SVs were performed in parallel.

Identification of sequence variants in the PAX6 transcriptional unit and regulatory region from WGS
WGS VCF files for all 51 individuals were filtered with the VEP-G2P plugin 21 using the EyeG2P 20 data set to detect high-impact and moderate-impact changes within genes known to cause genetic eye disease.Likely pathogenic PAX6 sequence variants identified, described in the following, are listed in table 2 along with the American College of Medical Genetics and Genomics (ACMG) pathogenicity classifications. 1 2 This revealed one proband (RPID 2134) with a novel frameshift variant (NM_000280.4PAX6: c.842_843insGT) affecting

Developmental defects
a coding exon of PAX6 constituting the sole 'false negative' for the prior screening of the PAX6 coding region (WGS data viewed in IGV in online supplemental figure S3).This case had been screened many years previously using denaturing HPLC analysis.
Although this was one of the 'trios', the 'paternal' sample was erroneously a duplicate of the maternal sample (short tandem repeat profiling in online supplemental table S3).We were thus unable to confirm whether the variant detected was de novo.

Essential splice site variants
Four families (three singletons, one trio) have heterozygous essential splice site (ESS) variants flanking exon 3, part of the PAX6 5′UTR.In RPID 877 and RPID 1019, both variants affect the 5′ base of intron 3 (IVS3+1, or c.-52+1), G>C and G>T, respectively (online supplemental figure S4A).Only the latter variant has been reported previously. 26A different variant at the same position (c.-52+1G>A) has been previously reported as resulting in skipping of exons 3, 4, 5 and 5a 27 ; exon 4 contains the translation start site.9][30] This variant occurred de novo in RPID 5645 (online supplemental figure S4B).The predicted effects of these sequence variants on splicing using the predictors in Alamut and SpliceAI are detailed in online supplemental table S4.Nested RT-PCR was performed on LCL-derived cDNA from RPID 1500 and showed evidence of abnormal splicing between exons 1 and 5 (figure 1B(i)).

Developmental defects
Variants in PAX6 CREs To identify causative CRE mutations at the PAX6 locus, we first created a BED file listing the GRCh38 genome coordinates of 35 previously characterised CREs 31 32 (online supplemental table S5).Intersecting this BED file with the annotated cohort VCF file identified 39 variants that passed quality filters and were present in at least 1 out of 52 sequenced individuals (online supplemental table S6).Of 35 CREs examined, 24 encompassed one or more variant (online supplemental table S5).Of 39 variants, 20 had allele count within the study population of nine or greater.
Of the remaining 19 variants, 15 had allele counts of one.Of 39 variants, 38 were present in gnomAD (online supplemental table S6), and the variant absent from gnomAD was a non-transmitted allele from an unaffected father in a trio. 23

Identification of large-scale SVs from WGS
A combination of direct inspection of candidate loci using the IGV 33 and genome-wide bioinformatic tools (Canvas 34 and Manta 16 ) was used to identify SVs from the available WGS data.A total of 17 different ultra-rare heterozygous SVs affecting PAX6 and FOXC1 were detected in 15 families (table 3 for genomic coordinates).
When compared with direct visual inspection, Canvas detected all deletions >10 kb at the PAX6 and FOXC1 loci (online supplemental figure S6) but not the two smallest PAX6 deletions (126 bp, RPID 75, and 1.36 kb, RPID 1524).Canvas also called the PAX6 deletion-duplication in RPID 1271 but was unable to detect the inversions in RPID 535 and RPID 774 as it uses read depth only.Manta detected all likely causative SVs but together with many false positive calls, so in practice these were identified solely by direct visualisation of the breakpoint regions using IGV.

Developmental defects
Balanced structural rearrangements disrupting PAX6 Two inversions of chromosome 11 were detected with breakpoints within or very close to PAX6 (figure 2, online supplemental figures S9 and S10).The PAX6 gene is directly disrupted in RPID 535, while in RPID 774 the breakpoint is between PAX6 and the critical CREs SIMO and HS5.

Complex structural rearrangements disrupting PAX6
RPID 356 was found to carry a de novo 6 kb heterozygous inversion with breakpoints at start of PAX6 in intron 4, with an adjacent 30 kb region of PAX6 deleted.In proband 1271, a 16 kb deletion encompassing the final six exons of PAX6 was associated with a 13 kb tandem duplication immediately 3′ of PAX6 (figure 3; also figure 2 and online supplemental figures S9 and S11).

Deletions encompassing FOXC1
Three heterozygous chromosome 6p deletions encompassing FOXC1 were identified in three probands: RPID 1142, RPID 1451 and RPID 1732 (table 3, online supplemental figure S12).These ranged from 33 kb to 83 kb in size.All three probands had aniridia; two out of three had glaucoma (one confirmed as congenital) and two out of three had congenital aortic or aortic valve anomalies (table 1).The combination of aniridia with congenital glaucoma and aortic valvular disease would be consistent with previously reported FOXC1 deletions. 36

Developmental defects
Breakpoint identification in a coincidental de novo reciprocal translocation t( RPID 356 has a de novo reciprocal translocation t(1,9)(p36.1;q22), which was detected by routine cytogenetic analysis following the clinical diagnosis of aniridia.Given that no PAX6 coding region mutation was identified on initial screening, the family was referred to our study to determine whether the breakpoint of this translocation could identify a novel locus or mechanism causing aniridia.However, as shown above, this individual has a second SV which disrupts PAX6 and explains the phenotype.Using IGV, discrepant paired-end reads mapped a single breakpoint on chromosome 1 and two different breakpoints on chromosome 9, consistent with a paracentric inversion on chromosome 9 and a reciprocal translocation with chromosome 1 (online supplemental figure S13).No clinical impact is suspected for these three breakpoints.

DISCUSSION
In purely diagnostic terms, short-read WGS has significant advantages over short-read whole exome sequencing (WES).First, WGS allows reliable analysis of the whole transcription unit of each gene.This power is evidenced by our identification of previously cryptic causative variants in the 5′UTR and deep intronic regions of PAX6 in 6 out of 22 (27.3%)diagnosed cases.The 5′UTR ESS variants perturb PAX6 splicing; however, consequential changes to the length of the PAX6 upstream ORF 37 and/or disruption of VAX2 binding 38 may also have mechanistic significance.More notably, we detected two deep intronic variants and tested the functional consequence of the novel one, predicted to result in an intron 8 donor site gain, using cDNA from an LCL derived from the proband.We could demonstrate the expected exon skipping 3′ to this variant, but we also found aberrant splice events 5′ to this intron, suggesting a more complex effect on splicing.While WGS may have better coverage of 5′UTRs than WES, it is particularly the deep intronic regions where it has a unique advantage.
A second advantage of WGS is more uniform per base coverage when compared with WES.This significantly improves our ability to detect disease-associated balanced structural variants (bSV) and CNVs.Our initial CNV screen was performed via direct inspection of the coverage depth change and unexpected pairing of end sequencing in proband BAM files using IGV. 33his proved to be the most diagnostically rewarding analysis undertaken in this study, yielding 15 of the 22 new diagnoses.Of these 15, 13 were CNVs (10 at PAX6 locus and 3 encompassing FOXC1) and 2 were balanced SVs (bSVs) with PAX6-disruptive breakpoints, the latter not easily detectable by bioinformatic SV calling with Manta due to noise.The high CNV yield 26 39 40 reflects both prescreening of the cohort for PAX6 coding variants and the historic nature of some samples in our cohort, as these anomalies would almost certainly be detected by modern high-resolution, array-based methods of copy number assessments now used in clinical diagnostic laboratories throughout the world.On the contrary, the two bSVs would be unlikely to be detected on standard clinical testing other than WGS.The identification of an apparently coincidental de novo balanced reciprocal translocation in RPID 356 is interesting but has been observed in other developmental disorders in which a second intragenic SV is subsequently determined to be causative. 41he third, and possibly most exciting, advantage of WGS in the diagnostic investigation of classic aniridia is the ability to identify causative cis-regulatory variants affecting the developmental expression of PAX6.CNVs and bSVs encompassing CREs of PAX6 but leaving the transcription unit intact have been recognised as resulting in functional haploinsufficiency for many years. 42 43Predicting the consequence of SNVs within CRE remains challenging, and currently only one de novo plausibly causative CRE SNV in classic aniridia has been reported. 44We did not identify any additional CRE SNVs in this study, although we did identify four SVs affecting only the PAX6 downstream regulatory region (three deletions and one bSV), leaving the gene itself intact.A similar PAX6 bSV is recently reported amongst a large, more diverse clinical diagnostic cohort 45 .
On the basis of the work from others 36 45 46 and ourselves, 47 the identification of FOXC1 deletions is not surprising from a human genetics perspective.There is remarkably little information about developmental genetic interactions between these two genes, although it has been shown that FOXC1 is a downstream direct target of PAX6 in the developing iris and ciliary body. 48 fourth strength of WGS is that it permits a search for new candidate genes, and the mechanisms of inactivating known genes, in unexplained cases.We did not identify any likely causative variants at loci other than PAX6 or FOXC1.A study with a larger number of trios would have greater power to detect new candidate loci.
We consider that the data presented here provide evidence that short-read WGS merits consideration as a primary investigation for classic aniridia.It certainly should be considered in cases with a normal array-based assessment of genome-wide copy number and PAX6 coding region sequencing.We are mindful that WGS analysis is not currently capable of explaining all cases of aniridia, and there remain 15 out of 37 families in this study in whom we have still not identified a causative variant.One useful emerging diagnostic technology is long-read nanoporebased genome sequencing.][51] 1μg genomic DNA was sheared by a Covaris ultrasonicator.The DNA fragments were assessed by gel electrophoresis, purified using AxyPrep Mag PCR clean-up kit (Axygen), incubated with End Repair Mix at 20°C for 30 min and then purified again.The repaired DNA fragments were A-tailed at the 3' end (37°C for 30 min), adapter ligated (16°C for 16 hr), purified and then size selected (350 bp) prior to PCR amplification and purification.
MGIEasy DNA library construction kits (MGI Tech) were used.
For Edinburgh Genomics: In addition to the QC outlined in the main manuscript, gDNA samples were further evaluated for quantity using Quant-iT Picogreen reagent (Thermo Fisher) and Spectramax BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s)  Green outer F (B1466) GCCAGTGAGGAGCGGTG

Pink inner R (B1680) TGTTGTGTCCCCATAGTCAC
BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s)

Figure 2
Figure 2 Structural variants (SV) identified on WGS affecting the wider PAX6 locus.Each SV is shown as horizontal bars (inv, inversion, teal; del, deletion, red; dup, duplication, green).The PAX6 topologically associated domain (TAD) is indicated by the Hi-C heatmap.The position of PAX6 cis-regulatory elements (CREs) is shown as track.Gene regulatory features such as the promoters and ATG are included.The position of PAX6 is shaded blue.Seven of the SVs have intragenic breakpoints and RPID 724 has a whole gene deletion.Three SVs are deletions of the downstream regulatory region, taking out CREs implicated in aniridia, notably SIMO and HS5.Similarly, the SV seen in RPID 774 inverts PAX6, disrupting its relationship with these enhancers.A smaller scale version of this figure, showing the full span of the largest SVs, is shown in online supplemental figure S9.RPID, research participant identifier; WGS, whole genome sequencing.

Figure 3
Figure 3Complex structural rearrangements of chromosome 11 in two unrelated individuals with aniridia.Aligned WGS data viewed with IGV, with reads viewed as pairs and coloured both by insert size and by pair orientation.Coordinates estimated from WGS data (GRCh38).(A) RPID 356: an individual with sporadic bilateral aniridia and congenital cataracts.Trio WGS data, including the unaffected parents, are shown in online supplemental figure S11.WGS data show a de novo 6 kb inversion involving the P1 promoter, all of the 3'UTR and the first coding exon (exon 4) of PAX6; next to this is a 30 kb deletion, which deletes the P0 promoter and several enhancers, including EE.The blue and teal colours both denote paired reads with abnormal pair orientation.In IGV, pair orientation is determined first, and only if this is as expected will abnormal insert size then be flagged.Therefore, the reads across this deletion are not flagged in red (as they would be in a simple deletion) as they also span the inversion.A drop in coverage depth is seen in the deleted area.(B) RPID 1271: an individual (single proband) with bilateral aniridia and cataracts.WGS data indicate a 16 kb deletion (red) involving the six last exons of PAX6 and a 13 kb tandem duplication (green) affecting the final intron of ELP4.Coverage depth is increased over the putative duplication and decreased over the putative deletion.IGV, Integrative Genomics Viewer; RPID, research participant identifier; WGS, whole genome sequencing.

XPS
Gemini plate reader (Molecular Devices, Wokingham, UK), and quality using an AATI Fragment Analyzer (Agilent, Stockport, UK) (total amount >1000ng and quality score >5 passed sample QC).gDNA samples were pre-normalised to a concentration range of 5-40ng/uL.Next generation sequencing libraries were prepared using an Illumina SeqLab specific TruSeq Nano High Throughput library preparation kit in conjunction with the Hamilton MicroLab STAR and Clarity LIMS X (4.2) Edition.200ng gDNA sample inputs were sheared to a 450bp mean insert size using a Covaris LE220 focused-ultrasonicator.The inserts were blunt ended, A-tailed, size selected, with TruSeq adapters ligated to each fragment before PCR amplification.Insert size for each library was evaluated using the Caliper GX Touch with a HT DNA 1k/12K/HI SENS LabChip and HT DNA HI SENS Reagent Kit, ensuring mean fragment sizes of 530bp to 950bp.The Roche LightCycler 480 and a Kapa Illumina Library Quantification kit were used to ensure the concentration of each library was between 5.5nM and 40nM.Libraries are normalised to 1.5nM and denatured for clustering and BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s) using a Hamilton MicroLab STAR with Genologics Clarity LIMS X (4.2) Edition.Libraries were clustered onto a HiSeqX Flow cell v2.5 on cBot2s and the clustered flow cell transferred to a HiSeqX for sequencing using a HiSeqX Ten Reagent kit v2.5.
allele frequency data in the VCF file.

Table 1 Continued
Studies Library.Protected by copyright.

Table 2
Likely pathogenic PAX6 sequence variants (NM_000280.4) ACMG/ACGS pathogenicity classification1(and posterior probability calculated using the Bayesian framework tool in DECIPHER 2 ).*Arguable whether PVS1 1 could be applied to this canonical splice site in a non-coding exon.ACGS, Association for Clinical Genomic Science; ACMG, American College of Medical Genetics and Genomics; CDS, coding sequence; del, deletion; ins, insertion; LP, likely pathogenic; mod, moderate; n/a, not available; P, pathogenic; RPID, research participant identifier; supp, supporting.Studies Library.Protected by copyright.on February 14, 2024 at ROYAL SCHOOL OF VETERINARY http://jmg.bmj.com/J Med Genet: first published as 10.1136/jmg-2023-109181 on 30 November 2023.Downloaded from No de novo CRE SNVs or INDELs were identified.It thus seems very unlikely that any of the CRE variants are of clinical significance for aniridia.Categorisation of de novo variants in genes other than PAX6The only individual with de novo SNVs or INDELs outwith PAX6 and with no causative SVs (described in the next section) was RPID 2469, who had five such variants (online supplemental table S7).This was the aforementioned trio for which the parents' DNA was extracted from LCLs.Only the variant in the gene encoding adenosylhomocysteinase 3 (AHCYL2: p.(Leu352Met)) merited further consideration.This variant (NM_015328.4(AH-CYL2):c.1054C>A) is not present in gnomAD, and has CADD and REVEL scores of 24.3 and 0.73, respectively.There is no known Mendelian disease-gene link for AHCYL2 and no claim can be made on the clinical significance of this variant.

Table 3
Likely pathogenic structural variants altering the PAX6 or FOXC1 loci 35his individual's structural variant was detected independently elsewhere and is patient 11 in the cited publication.35†ExonnumberingforPAX6 is as per NM_000280.4(-5a,13exons)DRR,Downstreamregulatoryregion;RPID,research participant identifier.Whole or partial deletions of PAX6Five individuals or families were found to have simple heterozygous deletions involving the PAX6 transcription unit (figure2, online supplemental figure S7): a whole gene deletion of 191 kb (RPID 724) and four partial deletions of 19 kb (RPID 1496), 1.4 kb (individual 1524), 0.13 kb (RPID 75) and 41 kb (RPID 2464).Each of these variants is expected to result in PAX6 haploinsufficiency.The variant detected in individual 1524 was subsequently found to have been identified independently by others.35DeletionsencompassingPAX6cis-regulatory domains RPID 1191, RPID 1361 and RPID 1647 were identified with likely causative deletions encompassing well-characterised CREs that control the developmental expression of PAX6 (figure2, online supplemental figures S8 and S9).

Table S2 :
Nested primers used in RT-PCR splicing analysis: oligonucleotide sequences

Table S3 :
Short tandem repeat (STR) genotype profiling of family 2134 confirming that the two parental samples were the same at the 10 STR loci tested, and indicating (as suspected on WGS) that the "paternal" sample was in fact a duplicate of the maternal sample

Table S6 :
Variants called within PAX6 CREsBMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s)

Table S7 :
De novo analysis results of trio whole genomes in the aniridia cohort Abbreviations: subs, substitution; del, deletion; dup, duplication; ins, insertion; IVS, intervening sequence/intron; CDS, coding sequence; P, pathogenic; U, uncertain; LP, likely pathogenic (all ACMG/ACGS criteria as previously) BMJ Publishing Group Limited (BMJ) disclaims all liability and responsibility arising from any reliance Supplemental material placed on this supplemental material which has been supplied by the author(s)