Summary
Autism Spectrum Disorder (ASD) exhibits an ∼4:1 male-to-female sex bias and is characterized by early-onset impairment of social/communication skills, restricted interests, and stereotyped behaviors. Disruption of the Xp22.11 locus has been associated with ASD in males. This locus includes the three-exon PTCHD1 gene, an adjacent multi-isoform long noncoding RNA (lncRNA) named PTCHD1-AS (spanning ∼1Mb), and a poorly characterized single-exon RNA helicase named DDX53 that is intronic to PTCHD1-AS. While the relationship between PTCHD1/PTCHD1-AS and ASD is being studied, the role of DDX53 has not been examined, in part because there is no apparent functional murine orthologue. Through clinical testing, here, we identified 6 males and 1 female with ASD from 6 unrelated families carrying rare, predicted-damaging or loss-of-function variants in DDX53. Then, we examined databases, including the Autism Speaks MSSNG and Simons Foundation Autism Research Initiative, as well as population controls. We identified 24 additional individuals with ASD harboring rare, damaging DDX53 variations, including the same variants detected in two families from the original clinical analysis. In this extended cohort of 31 participants with ASD (28 male, 3 female), we identified 25 mostly maternally-inherited variations in DDX53, including 18 missense changes, 2 truncating variants, 2 in-frame variants, 2 deletions in the 3’ UTR and 1 copy number deletion. Our findings in humans support a direct link between DDX53 and ASD, which will be important in clinical genetic testing. These same autism-related findings, coupled with the observation that a functional orthologous gene is not found in mouse, may also influence the design and interpretation of murine-modelling of ASD.
Autism Spectrum Disorder (ASD) refers to a group of neurodevelopmental conditions characterized by a heterogeneous early-onset impairment of social interaction and communication, combined with the presence of restricted, repetitive, and stereotyped behaviors and interests1–4. Although cognitive development is variable in affected individuals, these disorders are the leading contributor to disability in preschool-age children3; 5; 6. The underlying abnormalities in brain development and functional connectivity are presumed to be secondary to interactions between the genetic and environmental factors, making ASD a complex and multifactorial condition2; 7.
ASD can occur in association with pathogenic sequence-level variants and copy number variations (CNVs; also, structural variations, SVs) in specific neuronal genes8–10. In a recent comprehensive annotation of all publicly available ASD genome sequence data, 134 ASD-associated genes were identified and ∼15% of families studied had a clinically relevant rare variant. ASD-relevant gene lists are being curated by the Simons Foundation Autism Research Initiative (SFARI)11 and the Evaluation of Autism Gene Link Evidence (EAGLE)12 with both datasets available at https://gene.sfari.org.
Among the loci associated with ASD, chromosome Xp22.11 has been found to be disrupted in autistic males13–16. We present here a summary of the latest genotypic and phenotypic features mapping to this locus (Figure 1). It encompasses DDX53, PTCHD1 (MIM * 300828), and the long noncoding RNA (lncRNA) PTCHD1-AS (PTCHD1 Antisense RNA (Head-To-Head))15. Ptchd1 mutant mice show deficient attention and cognition, while variants in the human PTCHD1 gene are linked to intellectual disability, impaired vision and more rarely, autistic features (Autism X-linked 4, AUTSX4 - MIM # 300830)17–19. However, many deletions detected in autistic participants also disrupt exons of the upstream lncRNA PTCHD1-AS and/or the protein-coding gene DDX5315; 18; 20 leaving uncertainty of a strict genotype-phenotype correlation. In our previous work, deletion mapping of males with ASD indicated that the exon 2-4 region of PTCHD1-AS (encompassing DDX53) represents a critical region and mouse modeling reaffirms a role for the lncRNA (unpublished).
The Xp22.11 locus includes DDX53 (MIM * 301079), PTCHD1 (MIM * 300828), and the long noncoding RNA (lncRNA) PTCHD1-AS (PTCHD1 Antisense RNA (Head To Head)). A number of deletions identified in ASD participants were found to disrupt exons of the upstream lncRNA PTCHD1-AS and/or the DDX53 gene. a = SFARI Gene with curated EAGLE gene scores: https://gene.sfari.org/; b = ClinGen https://clinicalgenome.org/; c = EAGLE gene curation https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10357004.
DDX53 is an intron-less gene of 3,630 nucleotides located within intron 3 of the canonical transcript of PTCHD1-AS encoding a 631-amino acid DEAD-box RNA helicase21. Phylogenetic analysis suggests DDX53 and DDX43 are highly related and that DDX53 appears to be a retrogene derived from the 16-coding exon DDX43 progenitor on chromosome 622. The majority of retrogenes appear to originate from X-linked progenitors, possibly as a result of hypothesized compensation mechanisms23; 24. DDX43 (72.8 kDa) and DDX53 (71.2 kDa) proteins are also similar in size.
Moreover, DDX53 is also described as being a cancer-associated gene (CAGE), since it was first detected in human gastric cancer cell lines21. It is overexpressed in human cancers, where it promotes cancer stem cell-like properties such as self-renewal, cell motility, tumor spheroid formation, and anticancer drug resistance25; 26. DEAD-box RNA helicases have crucial roles in mRNA metabolism, serving as ATP-dependent helicases and unwindases, chaperones, and mediators of association or dissociation with protein interactors (RNPases)27; 28. Additionally, RNA helicases act as transcriptional co-activators or co-repressors, driving target mRNAs either to protein synthesis or degradation28. As such, these enzymes contribute to coordinate gene expression programs within the cell28.
In this study, we investigated 7 participants (6 males and 1 female) with ASD, with or without impaired psychomotor development, in whom 6 distinct maternally-inherited variants in DDX53 were identified. The study involving human participants was conducted according to the guidelines of the Declaration of Helsinki. Ethical review and approval were obtained locally through the Gaslini Children’s Hospital Research Committee for families I-VI (Comitato Etico della Regione Liguria, n.163/2018) and The Hospital for Sick Children Research Ethics Board. Family VI was consented at UCLA within IRB#15-000766 of the Undiagnosed Diseases Network. For all families, informed consent for the study and publication of research work was obtained from the participant or parents/ legal guardians of all the enrolled participants. After the identification of the index case, further individuals were included in the study cohort based on the identification of rare and potentially deleterious variants in DDX53, along with overlapping autistic and neurodevelopmental phenotypes. The initial participants were recruited through international collaboration involving clinical and research centers based in France, Italy, and the USA (further details available in the Supplementary Material), and also using GeneMatcher29. Enrolled participants were evaluated by pediatric neurologists, neuropsychiatrists, and neurogeneticists. For ASD assessment, thorough clinical examination and dedicated tests were employed, such as the Autism Diagnostic Interview-Revised (ADI-R) and Ritvo Autism Asperger Diagnostic Scale– Revised (RAADS–R) (Supplementary Material).
For genetic testing, trio-exome sequencing was performed for participant families I, II, and VI, and trio-whole genome sequencing was performed for VII. Participants #3, and #5 were investigated through the Next Generation Sequencing-based Autism/ID Xpanded Panel (GeneDx, Gaithersburg, USA) and participant #4 had targeted sequencing (GeneDx, Gaithersburg, USA) (Supplementary Material). Genomic DNA was extracted from peripheral blood lymphocytes of the probands and parents, with the sequencing being performed as described previously (Supplementary Material). Variants were filtered according to minor allele frequency ≤ 0.001 in Genome Aggregation Database (gnomAD), were assessed for levels of conservation (Genomic Evolutionary Rate Profiling – GERP), and functional impact was predicted according to different in silico tools (Supplementary Material), as well as the Variant Effect Predictor (VEP) pipeline from Ensembl.
Candidate variants in known disease genes were classified according to ACMG guidelines30. Variants were mapped onto the protein structures and intolerance to variation of the affected amino acid residues was analyzed using the Metadome online software (https://stuart.radboudumc.nl/metadome/method)31. Copy number variation (CNV) assessment was performed either through array comparative genomic hybridization (aCGH) or whole genome/exome-based CNV calling (Supplementary Material). The DECIPHER database and Database of Genomic Variants were used to interpret detected CNVs32; 33. DDX53 variants are reported according to the NM_182699.4 transcript, corresponding to the unique isoform of the DDX53 protein NP_874358.2 (https://www.ncbi.nlm.nih.gov/gene/168400).
We identified six distinct variants in DDX53 in the reported individuals: c.871A>G, p.(Arg291Gly) in #1; c.834G>A, p.(Met278Ile) in #2; c.567_568insTGCAGGT, p.(Glu190Cysfs*3) in #3 and #4; c.977G>A, p.(Ser326Asn) in #5; c.1736C>T, p.(Ala579Val) in #6; c.259G>T, p.(Gly87Trp) in #7. All variants were maternally inherited and segregated with the phenotypes in all families (Figure 2A). In family III, both participants harbored the same p.(Glu190Cysfs*3) variant, although no information about a potential non-skewed X-inactivation in the female sibling (#4) was available. In family IV, the p.(Ser326Asn) variant identified in the proband (#5) was mosaic (4%) in the mother. In family V, the brother of participant #6 was reported to show autistic features, but segregation analysis could not be performed. In family VI, there were no other maternal male relatives for further segregation analysis available. All of the detected DDX53 variants are rare (maximum allele frequency 0.000005534) and absent in the hemizygous state in gnomAD. The missense variants p.(Gly87Trp), p.(Met278Ile), p.(Arg291Gly), p.(Ser326Asn), and p.(Ala579Val) affect conserved residues (GERP scores ranging between 3.55 and 4.59) within or in close proximity to the KH, DEXDc, and HELICc domains (Figure 2B).
Pedigrees of the six families showing the segregation of the DDX53 variants in affected individuals and the parents. All variants were inherited from unaffected mothers. The carrier status is indicated by small-filled circles. The mosaic status of the mother of the proband #5 is indicated by a small empty circle. B) Schematic representation of the total of the DDX53 variants mapped to the unique protein isoform (NP_874358.2). The variants identified in the described families and those identified in participants from the SFARI and MSSNG databases are reported in blue and black, respectively. DEXDc = DEAD-like Helicases superfamily domain; HELICc = helicase conserved C-terminal domain; KH = K homology domain.
These changes are predicted to be damaging according to in silico tools (Table S1). The frameshift variant p.(Glu190Cysfs*3) is likely to lead to either nonsense-mediated mRNA decay (NMD) or the formation of a truncated transcript. Pathogenicity using the ACMG classifications could not be definitively assigned for this gene, as ACMG guidelines are only applicable for genes with already established gene-disease associations. The predicted effects of these variants are indicated in Tables 1 and 2.
The DDX53 participants included in our cohort presented with marked ASD-related features, including poor social interaction, limited communication skills, self-stimulating behaviors, weak visual problem-solving skills, and sensory hyperresponsiveness (Table 1). Eye contact avoidance was observed in all cases but #7. Motor and/or verbal stereotypies were present in four participants (#3, #5, #6, and #7), sometimes in association with ritualistic behaviors (Table S2). A form of high-functioning autism within the spectrum of Asperger syndrome was diagnosed in #2. Interestingly, while two participants from Family III presented with similar overlapping ASD phenotypes, female participant #4 showed a more severe phenotype including significant speech limitation, social avoidance, sensory hypersensitivity, and impaired visual problem-solving skills (no additional variants explaining this were found). Autistic features were associated with impaired psychomotor development in five out of seven participants (#3-#7), variably affecting speech development and/or motor abilities. However, this developmental delay only led to variable skills, generally in the mild to moderate range, in participant #7. Additional neuropsychiatric manifestations consisted of seizures (#1, #6, and #7), heterogeneous sleep disorders (#3-#7), and behavioral comorbidities (#3, #5-#7), such as attention-deficit hyperactivity disorder (ADHD), anxiety, and aggressiveness.
After the identification of the DDX53 variants in the clinical cohorts, we searched the SFARI (https://www.sfari.org/resource/sfari-gene/) and MSSNG (https://research.mss.ng) databases. Within SFARI11; 34, we screened for DDX53 variants in the Simons Simplex Collection (SSC) and Simons Foundation Powering Autism Research (SPARK) datasets. SSC consists of WGS data from 2,600 simplex families, each containing one proband with ASD as well as unaffected parents and siblings. SPARK (WES1, WES2) contains the whole exome sequence for 43,259 participants (including 21,900 with ASD) with phenotype information. Additionally, we screened the MSSNG WGS database8; 35–37, which at the time consisted of 11,312 individuals, including 5,100 with ASD.
By screening the SSC, SPARK, and MSSNG cohorts, we detected 19 additional interesting variations in DDX53 in 24 individuals, including 22 males and 2 females, presenting with mostly isolated ASD. These included 13 missense changes, 1 stopgain, 2 in-frame variants, 2 distinct deletions within the 3’UTR and one deletion CNV (Table 2). Fourteen of these variations were of maternal origin, whereas information about the segregation was not available in the remaining cases. Resembling those identified in our cohort, the DDX53 variants detected in participants from the open-science autism datasets are exceptionally rare (max allele frequency of 0.0000563) and are absent in hemizygous state in gnomAD. Most missense changes affect conserved residues within different functional domains of the DDX53 helicase (Figure 2B) and are predicted deleterious by in silico tools (Table S3). The novel truncating variant p.(Trp8*) was predicted to exert a deleterious effect on protein expression, whereas the in-frame changes (p.(Gln591del) and p.(Met381_Leu382delinsIleMet) resulted in the deletion or substitution of highly conserved amino acids, respectively. Of note, in the SPARK dataset, we also identified participants that were already part of our main cohort: subjects #3 (SP0039198), #4 (SP0042895), and #6 (SP0090623).
Deletions within the 3’UTR of DDX53 were identified in five participants. The [GRCh38] Xp22.11(23002082-23002458)x1 was detected in one individual (7-0111-003) and the [GRCh38]Xp22.11(23002148-23002522)x1 in four affected brothers (1-0627-003, 1-0627-005, 1-0627-006, and 1-0627-007). These deletions selectively involve a portion of the 3’UTR of the gene and are absent in the DECIPHER database. Overall, the identified DDX53 variations are predicted to affect protein function through the alteration of conserved residues within functional domains or the loss of a functional transcript/part of the 3’UTR. Additionally, a maternally inherited, 21.8 kb deletion [GRCh38]Xp22.11(22987677-23009507)x1 impacting only DDX53 and not exons of PTCHD1-AS was observed in one affected female. However, this individual also carries a 331 kb de novo deletion [GRCh38]7q11.22 (70441029-70771836)x1 impacting 5 exons of AUTS2 (MIM*607270), a gene strongly associated with ASD38. In the aforementioned individual, the AUTS2 deletion is likely involved in the ASD, but the carrier status of the DDX53 deletion may also need to be considered in male offspring.
In total, in our extended cohort of 31 individuals, including 28 males and 3 females, we identified 25 distinct DDX53 variations. These included 18 missense changes, 2 truncating variants, 2 in-frame variants, 2 deletions of the 3’UTR, and one complete gene deletion. Four variants were found in more than one individual, including p.(Glu190Cysfs*3) in two (#3 - SP0039198 and, #4 - SP0042895); p.(Asp575Gly) in two (SP0126624 and 1-0551-003); and p.(Gln591del) in two (SP0066816 and SP0066817). Of note, the [GRCh38]Xp22.11(23002148_23002522)x1 deletion was also identified in four participants from the same family with a similar deletion found in another family with one affected male. Missense changes affected conserved residues within or in proximity to the KH, DEXDc, or HELICc functional domains (Figure 1B). In silico tools revealed that these variants are variably deleterious (Table S1 and Table S3) and mostly map to amino acid residues intolerant to variation according to Metadome (Figure 3 and Table S4). One of the in-frame variants included the indel p.(Met381_Leu382delinsIleMet), causing the substitution of two conserved amino acids (GERP scores = 4.3) within the DEXDc domain, and p.(Gln591del), resulting in a deletion of a single conserved residue (GERP score = 4.46) in the C-terminus of the protein. The two truncating variants, which were identified in three participants, featured the stopgain p.(Trp8*) and the frameshift p.(Glu190Cysfs*3). These variants are likely to be deleterious based on the predicted severe impact on protein function and the loss of function intolerance of DDX53 (probability of loss intolerance (pLI) of 0.66 and a loss of function observed/expected upper bound fraction (LOEUF) of 0.847). Of note, affected females (#4 - SP0042895, SP0003053 and 1-0485-003) either harbored truncating DDX53 changes or a whole gene deletion. Although further studies are required to elucidate the X-inactivation pattern in these cases, this observation suggests that a severe DDX53 loss of function may lead to a clinical phenotype in female carriers.
Using the Metadome software (https://stuart.radboudumc.nl/metadome/), we mapped the variants identified in our cohort and in the SFARI and MSSNG datasets to the DDX53 protein. Most variants were found to affect amino acid residues that showed intolerance to variation according to their variation in the gnomAD dataset. While some amino acids were rarely impacted by genetic changes, variants in other residues are absent in gnomAD.
We scrutinized the mouse databases for Ddx53 and based on the available mouse reference data at UCSC Genome Browser (GRCm39/mm39), there does not appear to be an orthologous transcript in the expected syntenic region (Figure 4)39. Instead, there is a gap in alignment between the mouse and human regions; however, that gap does not exist in the DNA of higher mammals.
A) Graphical representation of human, mouse, and rat syntenic regions and breaks in chrX:10,445,310-35,803,753 (GRCh38). Syntenic alignment between human and mouse show inverted orientation of two corresponding segments between human and mouse. Syntenic alignment between human and rat is shown to highlight the mouse-specific inversion of this chromosomal segment, and the shared syntenic break located just upstream of the PTCHD1 gene. B) Inset of human-mouse syntenic region at the PTCHD1-AS locus. C) Full phylogenetic tree of DDX53 and DDX43 (right panel) created by Treefam. Red triangle denotes a duplication event. Green circle denotes a speciation event. Inset of DDX53 orthologs (left panel). D) Phylogenetic tree of DDX53 and DDX43 in common model organisms.
We designed primers in the repeat-free sequence on each side of the alignment gap to determine if there was any additional sequence in this gap not captured in the murine reference genome.
(forward: ttcaacctatttgtgtgtgcatc; reverse: tccagcacagtattaccagaaatagt). Using BAC clone RP23-405B14 from the C57BL/6 mouse strain as DNA template, PCR was performed with an expected product size of 4,029 bp. Based on gel electrophoresis the product appeared to be ∼4 kb. This product was cloned into a Topo TA cloning vector (Invitrogen), and Sanger sequenced with the original primers, as well as walking primers, to generate the entire 4,029 bp cloned product. The Sanger sequence obtained corresponds with the reference sequence GRCm39/mm39. We repeated this PCR using template DNA from mouse genomic DNA (purchased from Cedarlane) and DNA extracted from internal C57BL/6 mouse brain P14. In each case, the expected product size was amplified and the Sanger sequencing agreed with the reference sequence. All data suggest that there is no additional sequence in this region, in this strain, and therefore a mouse Ddx53 gene does not appear to be present in the syntenic region to DDX53. These observations are further validated because no obvious RNA transcripts were found for Ddx53 in any database nor in our unpublished murine data. The rat ortholog (Ddx43), however, is annotated as a DDX43/DDX53 homolog and appears to be functional (https://www.genenames.org/tools/hcop/#!/). The naked mole-rat model has a more specific DDX53 homolog (https://www.ncbi.nlm.nih.gov/gene/101703948/). Other potentially functional homologs are present in rabbits and other small mammals (https://www.ncbi.nlm.nih.gov/gene/168400/ortholog/?scope=9347&term=DDX53) (Figure 4)40.
RNA helicases contribute in mRNA metabolism, mediating a range of biochemical activities not restricted to RNA unwinding, such as ribosome biogenesis, RNA degradation, and splicing41. While processive RNA helicases function by translocating along the nucleic acids, the non-processive helicases selectively unwind local secondary structures without translocating on their substrates28; 42. In humans, the two largest families of helicases are: the DEAD box (DDX) family, encompassing 41 members (https://www.genenames.org/data/genegroup/#!/group/499); and the DEAH box (DHX), including 16 members (https://www.genenames.org/data/genegroup/#!/group/500). The DEAD-box helicases belong to the non-processive group43. All DDX helicases share a conserved core consisting of two tandem RecA domains, with recognizable sequence motifs involved in RNA binding, and ATP binding and hydrolysis41; 42. Additional C- and N-terminal extensions vary among the different members and consist of auxiliary domains implicated in RNA recognition and the regulation of ATPase activity41;42.
Genetic variants affecting the function of DDX helicases have been shown to be involved in other neurodevelopmental disorders44. In particular, de novo variants in DDX3X (MIM * 300160), mapping to chromosome Xp11.3-p11.23, cause a syndromic neural condition (MIM # 300958) with dysmorphism, congenital defects, and neurobehavioral comorbidities including autistic features, either in males or females with different disease severity45; 46. More recently, although in absence of supporting functional evidence, missense changes in DDX54 (MIM * 611665) have been associated with psychomotor development impairment in three unrelated participants, of whom one exhibited repetitive movements and behaviors44.
DDX53 is a little characterized RNA helicase 21 whose main functional core consists of the DEAD-like Helicases superfamily (DEXDC) domain, which contains the ATP binding site and mediates the ATP-dependent RNA unwinding activity21; 47. Additionally, the protein has two auxiliary domains; the N-terminal K homology (KH) domain, which favors single-strand RNA recognition and binding, and the helicase conserved C-terminal (HELICc) domain, which interacts with the DEXDC domain for the helicase activity41; 47. So far, the DDX53 helicase function has only been elucidated in cancer, where its overexpression appears to mediate the acquisition of oncogenic properties in tumoral cells21; 25; 26. Based on its elevated gene expression in testis, polymorphisms in DDX53 may be associated with azoospermia, but this putative finding requires further testing48; 49.
It is noteworthy that in our own recent study, Neurogenin 2 (NGN2)-derived cortical excitatory neurons harboring termination codons in DDX53 showed a comparable activity in the multi-electrode array to isogenic control lines50. The lack of a distinctive electrophysiological phenotype and the ability of the knockout induced pluripotent stem cell (iPSC) lines of generating excitatory cortical neurons suggested that loss of DDX53 does not lead to a glutamatergic dysregulation phenotype, which may argue against an overt role in ASD50. However, the isogenic iPSC-derived control neurons investigated in this study showed low levels of DDX53 transcripts, resulting in protein levels below the threshold for western blot detection50. Thus, it is possible that these cells are not the most appropriate model to investigate DDX53 function; a conjecture further supported through the expression of the DDX53 gene in the human brain, where significant RNA expression is limited to the cerebellum15 (https://www.proteinatlas.org/ENSG00000184735-DDX53/tissue). The cerebellum is responsible for the maturation of nonmotor neural circuitry, thus having a crucial role in guiding cognitive development51. Furthermore, specific cerebellar subzones modulate the activity of neocortical substrates contributing to the determination of social interaction patterns during developmentally sensitive periods51.
In summary, our genetic data suggest a role for DDX53 in ASD. Interestingly, at the same Xp22.11 locus, there is also supporting evidence based on non-overlapping microdeletions, for the PTCHD1-AS lncRNA, which encompasses DDX53, to be independently involved in ASD15; 20 (Figure 1). Moreover, at the same locus, for the protein-coding PTCHD1, non-synonymous variants have also been found to cause intellectual disability, and in a few individuals, an accompanying diagnosis of autism was present18. In murine models, Ptchd1 deficiency leads to cognitive dysfunction and abnormal behavior with significant hyperactivity; however, despite these behavioral abnormalities they do not strictly resemble the autism spectrum17; 19. Ptchd1-as cell20 and mouse models (unpublished) are being generated. As additional participants with ASD that have sequence-level or deletion variants impacting DDX53, PTCHD1-AS and/or PTCHD1 are identified, we anticipate an increasingly complex social-behavioral and clinical presentation when two or more of these co-located genes are impacted. Given the newly found role for DDX53 in ASD, and that there is no apparent functional ortholog (Ddx53) in mouse, generating a ‘humanized’ Ddx53 murine model may be important for future mouse-modelling of autism.
Data Availability
The data that support the findings of this study are available from the corresponding authors upon request or included in the databases mentioned and linked in the manuscript.
Funding
Funding was provided by the University of Toronto McLaughlin Centre, Autism Speaks, Autism Speaks Canada, Ontario Brain Institute, SickKids Foundation and the Italian Ministry for Education, University and Research (Ministero dell’Istruzione, dell’Università e della Ricerca - MIUR) PRIN2020 code 20203P8C3X (Alfredo Brusco). Additional funding was partially provided by NIH award U01NS134356 and supported by the California Center for Rare Diseases within the Institute for Precision Health at UCLA. S.F.N holds the Dr. Allen and Charolotte Ginsburg Endowed Chair in Translational Genomics at the David Geffen School of Medicine. S.W.S. holds the Northbridge Chair in Pediatric Research at The Hospital for Sick Children and the University of Toronto.
Author Contributions
Conceptualization: MS, SWS; Formal Analysis: MS, JLH, NBS, CS, MSR, JRM, SWS; Funding Acquisition: MS, SFN; SWS; Investigation: MS, CAB, JLH, BT, CS, MSR, JRM, SYK, FZ, SWS; Methodology: MS, SWS; Project Administration: MS, JLH, SWS; Resources: MS, PWF, LG, GA, VP, AB, RK, SP, HFP, LL, PPM, IH, SVM, ED-B, BER, SFN, SWS; Supervision: MS, SWS; Visualization: MS, JLH, CS, MSR, SWS; Writing: MS, JLH, NBS, CS, JRM, SWS
Web resources
CADD; https://cadd.gs.washington.edu
ClinVar; https://www.ncbi.nlm.nih.gov/clinvar
Combined Annotation Dependent Depletion (CADD); http://cadd.gs.washington.edu
Database of Genomic Variants (DGV); http://dgv.tcag.ca/dgv/app/home
DECIPHER; https://decipher.sanger.ac.uk
Gene Cards; https://www.genecards.org
Gene Matcher; http://www.genematcher.org
Genome Aggregation Database (GnomAD); http://gnomad.broadinstitute.org
Genomic Evolutionary Rate Profiling (GERP); http://mendel.stanford.edu/SidowLab/downloads/gerp/
gnomAD; https://gnomad.broadinstitute.org
Metadome online software; https://stuart.radboudumc.nl/metadome/method
MSSNG; https://research.mss.ng
Online Mendelian Inheritance in Man; https://www.ncbi.nlm.nih.gov/Omim
Protein Atlas; https://www.proteinatlas.org/ENSG00000184735-DDX53/tissue
Polyphen-2; http://genetics.bwh.harvard.edu/pph2/
PubMed; https://www.ncbi.nlm.nih.gov/pubmed
RefSeq; https://www.ncbi.nlm.nih.gov/refseq
SFARI; https://www.sfari.org/resource/sfari-gene/
SIFT; https://sift.bii.a-star.edu.sg
UniProt; https://www.uniprot.org
UCSC Human Genome Database; https://www.genome.ucsc.edu
Varsome; https://varsome.com
Variant Effect Predictor (VEP) - Ensembl; https://www.ensembl.org/info/docs/tools/vep/index.html
Competing interests
At the time of this study and its publication, S.W.S. served on the Scientific Advisory Committee of Population Bio. Intellectual property from aspects of his research held at The Hospital for Sick Children are licensed to Athena Diagnostics and Population Bio. These relationships did not influence data interpretation or presentation during this study but are disclosed for potential future considerations. SVM is an employee of GeneDx, LLC. HFP is on the research advisory boards and speaker bureau for Takeda Pharmaceutical, AvroBio, Amicus Therapeutics, Sanofi, Alexion Therapeutics, Denali Therapeutics and Acer Therapeutics. All other authors declare no conflict of interest.
Table S1. Extended in silico analysis of DDX53 variants in our cohort.
Table S2. Extended clinical phenotype of participants with DDX53 variants.
Table S3. Extended in silico analysis of all DDX53 variants of participants from SFARI and MSSNG databases.
Table S4. Details of intolerance to variation for the amino acid residues affected by DDX53 variants.
Table S5. UDN membership list.
Acknowledgements
We acknowledge technical support from Bhooma Thiruvahindrapuram, and Sylvia Lamoureux. We thank the research participants and their families (including those participating in MSSNG) as well as the generosity of the donors who support this program. We acknowledge the resources of Autism Speaks and The Centre for Applied Genomics. We are grateful to all of the families described in this paper including those involved in the Autism Speaks MSSNG Project, as well as the SSC and SPARK sites including the principal investigators (A. Beaudet, R. Bernier, J. Constantino, E. Cook, E. Fombonne, D. Geschwind, R. Goin-Kochel, E. Hanson, D. Grice, A. Klin, D. Ledbetter, C. Lord, C. Martin, D. Martin, R. Maxim, J. Miles, O. Ousley, K. Pelphrey, B. Peterson, J. Piggot, C. Saulnier, M. State, W. Stone, J. Sutcliffe, C. Walsh, Z. Warren, and E. Wijsman). We appreciate obtaining access to genetic and phenotypic data on SFARI Base.
Abbreviations
- ACMG-AMP
- American College of Medical Genetics and Genomics and the Association for Molecular Pathology
- ADHD
- attention deficit-hyperactivity disorder
- ASD
- autism spectrum disorder
- ES
- exome sequencing
- iPSC
- induced pluripotent stem cell
- NMD
- nonsense-mediated mRNA decay
- SNVs
- single nucleotide variants
- CNV
- copy number variant
- SV
- structural variant
- SFARI
- Simons Foundation Autism Research Initiative
- EAGLE
- evaluating evidence for ASD