1q21.1 microduplication: large verbal–nonverbal performance discrepancy and ddPCR assays of HYDIN/HYDIN2 copy number

Microduplication of chromosome 1q21.1 is observed in ~0.03% of adults. It has a highly variable, incompletely penetrant phenotype that can include intellectual disability, global developmental delay, specific learning disabilities, autism, schizophrenia, heart anomalies and dysmorphic features. We evaluated a 10-year-old-male with a 1q21.1 duplication by CGH microarray. He presented with major attention deficits, phonological dysphasia, poor fine motor skills, dysmorphia and mild autistic features, but not the typical macrocephaly. Neuropsychiatric evaluation demonstrated a novel phenotype: an unusually large discrepancy between non-verbal capacities (borderline-impaired WISC-IV index scores of 70 for Working Memory and 68 for Processing Speed) vs. strong verbal skills – scores of 126 for Verbal Comprehension (superior) and 111 for Perceptual Reasoning (normal). HYDIN2 has been hypothesized to underlie macrocephaly and perhaps cognitive deficits in this syndrome, but assessment of HYDIN2 copy number by microarray is difficult because of extensive segmental duplications. We performed whole-genome sequencing which supported HYDIN2 duplication (chr1:146,370,001-148,590,000, 2.22 Mb, hg38). To evaluate copy number more rigorously we developed droplet digital PCR assays of HYDIN2 (targeting unique 1 kb and 6 kb insertions) and its paralog HYDIN (targeting a unique 154 bp segment outside the HYDIN2 overlap). In an independent cohort, ddPCR was concordant with previous microarray data. Duplication of HYDIN2 was confirmed in the patient by ddPCR. This case demonstrates that a large discrepancy of verbal and non-verbal abilities can occur in 1q21.1 duplication syndrome, but it remains unclear whether this has a specific genomic basis. These ddPCR assays may be useful for future research on HYDIN2 copy number.


INTRODUCTION
Microduplication of 1q21.1 is observed in~0.03% of adults. 1,2 It has a variable and incompletely penetrant phenotype including diverse dysmorphic features, macrocephaly, congenital heart disease (tetralogy of Fallot), connective tissue abnormalities and neuropsychiatric disorders including learning disabilities, developmental delay, autism spectrum disorders (ASD), attention deficit hyperactivity disorder (ADHD) and schizophrenia. [2][3][4][5][6][7][8] We describe here a 10-year-old boy with a 1q21.1 microduplication with a predominantly neuropsychiatric and cognitive phenotype. In contrast with the low IQ that is frequently observed, he had a previously-unreported feature: a striking discrepancy between verbal (superior range) and non-verbal (borderline) cognitive performance.
Following a comprehensive medical and neuropsychiatric evaluation, we carried out whole-genome sequencing (WGS) to search for a genomic basis for the phenotype. We also developed droplet digital PCR (ddPCR) assays for copy number (CN) of HYDIN2 (1q21.1) and its 16q22.2 paralog HYDIN, because HYDIN2 was previously hypothesized to play a role in changes in brain size (normal in this case) and other neuropsychiatric features. 5 HYDIN2 is classified as a pseudogene but is expressed in CNS. 5,9 Most 1q21.1 CNVs form between segmental duplication (SD) regions BP2-BP3 (thrombocytopenia-absent radius syndrome; TAR) or the larger, more distal BP3-BP4 region (producing neuropsychiatric and other features). 3,10 Typical distal boundaries arẽ 144.6-146.3 Mb (hg18) 4,7 (hg19: 145. 8-147.8).

CASE REPORT
T is a boy with two siblings who were in good health. The primary guardian provided written informed consent. There is no known parental consanguinity, and no known maternal family history of psychiatric disorder or learning disability. Father was unavailable. T was born by vaginal delivery at 38 weeks + 5 days of amenorrhea. Pregnancy was uncomplicated, without neonatal resuscitation or jaundice. Mother was 23 and father 27 at conception. Birth weight was 3480 g, length 49 cm and head circumference 35 cm (all normal). Apgar scores were 10/10 (1/5 min). At age 10, height was 138 cM (50 th percentile), weight was 30 kg (below the 25 th percentile) and head circumference was 54 cm (+0.5 SD). He was examined by a medical geneticist at age 5½ for language delay. A comprehensive neuropsychiatric evaluation for behavioral problems was conducted beginning at 9 years-11 months.
Developmental history He walked at 24 months and had difficulty with fine motor skills (handwriting, drawing), coordination (ongoing difficulty tying shoes, buttoning clothing, riding a bicycle), speech and language delay (first words~18 months; combinations of words~3 years) and articulation (e.g., "totolat" instead of "chocolat"). He was toilet trained by age 6. He began speech and occupational therapy at age 4. At 10 he was attending a therapeutic school for learning disabilities. The school described him as a non-reader with articulation problems but with sustained oral expression. He had trouble tracking time, was anxious and impulsive with poor attention and memory, and was socially isolated.

Medical history
As an infant he had repeated bronchitis and ear infections, treated with tympanostomy tubes. At age 7½ he developed difficulty falling asleep and repeated awakenings every 2-3 h. Physical examinations of the heart, lungs, abdomen and eyes were repeatedly normal. No genito-urinary abnormalities were recorded (no structural evaluations were completed).

Morphological features
At ages 5½ and again at 10 he was described as having a long trunk, a long, square face, widely spaced eyes (hypertelorism) and large cheeks. Philtrum was long and well defined, upper lip was thin, with a flattened nose and without the normal protrusion of the middle facial skeleton. His first toe was much shorter than his second, with a large gap between them. Fingers were also very short. Performance at age 10 consistent with child at age 6  WISC -IV Wechsler intelligence scale for children -IV 27 ADI-R autism diagnostic interview-revised 28 TEA-Ch test of everyday attention for children 29 L2MA spoken language, written language, memory, attention 30 N-EEL new tests for language assessment 31 ELO oral language assessment 32 NEPSY developmental NEuropsychological assessment 33 UDN-II construction et utilisation du nombre-2 34 SD standard deviation DA developmental age M-ABC movement assessment battery for children 35 EMG evaluation de la motricité gnosopraxique distale 36 Bender visual-motor test 37 Rey-A Rey-Osterrieth complex figure test 38 Novel 1q21.1 microduplication phenotype J Xavier et al.
Cognitive and neuropsychiatric assessment Clinical interview revealed a hyperactive (moving constantly on his chair) and impulsive boy with poor attention. He attempted to communicate but was hindered by phonological and articulatory impairments; at times he had excessive and disorganized speech, anxiety and emotional liability. He drew crude and violent fantasy scenes, saying that this calmed him down. He easily became frustrated and destroyed his drawings. His thinking was sometimes non-linear and disconnected with reality. He denied hallucinations when awake, but described difficulty falling asleep while seeing little animals trying to penetrate his head (possible hypnagogic phenomena). Neurological examination was normal except for absent achilles and patellar reflexes. EEG while awake at 4 years-8 months revealed normal auditory and visual evoked potentials; slowing (predominantly on the right), with theta waves; and frontal monomorphic delta waves without paroxysmal activity. EEG at age 10 showed rare slow temporal theta waves, sometimes sharp. Prolonged waking and sleep EEG at 10 years-3 months revealed isolated sharp waves (while awake), but no continuous spike waves during sleep.
Cognitive assessment Table 1 The most striking feature was the discrepancy between verbal and non-verbal performance (Verbal Comprehension Index in the superior range, with borderline-impaired Working Memory and Processing Speed). Full-scale IQ computation was invalid because of the degree of heterogeneity. ASD symptoms were mild (subthreshold), with only the Developmental domain in the abnormal range. Notable abnormalities were seen for attention, capacity for inhibition, phonological skills, short-and long-term memory impairment, and motor skills.

Neuropsychiatric diagnoses
Final ICD-10 diagnoses included ADHD combined type, mixed specific developmental disorders including a prominent phonological deficit, and developmental coordination disorder with visual-motor impairments. The presentation resembled multiple complex developmental disorder 11,12 (emotional regulation difficulties, impaired cognitive processing, confusion between reality and fantasy life, impaired social behavior and sensitivity) and multi-dimensional impairment. 13 Clinical genetic analysis FISH probes showed no abnormalities (chromosomes 15q11, 17p11.2, 22q11 and 22q13). Fragile X syndrome was excluded by FMR1 PCR testing (exon 1). Comparative genomic hybridization (Agilent 2 × 105 array) detected an interstitial microduplication in distal 1q21.1 (144,940,840-146,290,831 bp, hg18), confirmed by FISH (BAC 533N14, Kreatech). Mother did not carry the duplication.

MOLECULAR ANALYSIS
Because this patient had preserved verbal ability and normal head circumference, we wanted to establish definitively whether HYDIN2 was duplicated. HYDIN2 is a human-specific gene created by an incomplete duplication of the HYDIN ancestral gene~3.2 mya. 9,14 HYDIN2 lacks the first five and last two HYDIN exons. 9 The sequence homology (Fig. 1) leads to incorrect CNV calls over HYDIN due to 1q21.1 CNVs, 5 and the flanking SDs make it difficult to resolve CNV boundaries.
Whole-genome sequencing and CNV analysis We carried out 2 × 151 bp (350 bp insert) Illumina whole-genome sequencing (WGS) and obtained 42× genome coverage. After removal of PCR duplicates, CNV analysis based on sequencing read-depth 15 was implemented on hg38 alignment data using the union of calls from CNVnator 16 and ERDS. 17 A 2.21 Mb duplication was detected in 1q21.1 (chr1: 146,370,001-148,580,000, hg38, by WGS (Fig. 1a). We found no other rare CNVs with known or likely neuropsychiatric consequences and also no loss-of-function mutations in HYDIN2 or other apparent functional relevance to Patient T's phenotype (see Supplementary Methods for details of sequencing and of CNV and SNV/indel calling and annotation). Comparison of HYDIN2 vs. HYDIN sequence identified 1 kb and 6.1 kb "insertions" as the largest HYDIN2-specific segments. To estimate HYDIN2 CN, we determined that 573 of T's sequencing reads mapped to the 1 kb insertion vs. 369 in HapMap DNA NA12878 sequenced at 41×. The ratio of 185 suggested a heterozygous HYDIN2 duplication. ddPCR copy number assays of HYDIN and HYDIN2 We developed ddPCR assays to measure these CNs more rigorously. 18,19 We designed Taqman probes and corresponding forward and reverse primers targeting the 1 kb and 6.1 kb HYDIN2 unique regions and a 154 bp segment of HYDIN outside the HYDIN2 overlap region (Fig. 1b). The three ddPCR assays were successfully designed to quantify CN relative to reference gene RPP30, using a previously described algorithm. 20 Performance of these assays was validated (Table S1 and S2) using 43 samples from the independent Molecular Genetics of Schizophrenia cohort 7 with and without Affymetrix 6.0 microarray evidence for 1q21.1 CNVs (originally called with Birdsuite 2.0 21 ). (The two HYDIN2 ddPCR CN assays were always concordant.) • Twelve subjects had microarray calls of heterozygous distal 1q21.1 CNVs (7 duplications and 4 deletions in individuals with schizophrenia; 1 deletion in a control), of which 10 also had CNV calls including HYDIN on 16q22.2. By ddPCR, all had HYDIN CN = 2. For HYDIN2, 10 had CN concordant with the microarray calls; 1 had CN = 2 which was consistent with the start position of the microarray-based duplication (chr1:144,943,150 bp, hg18), distal to the location of HYDIN2; and 1 deletion case had CN = 2 by ddPCR, but a sample swap was established by comparing genotypes from the microarray vs. from WGS.
• Seven (5 schizophrenia cases, 2 controls) had CNV calls in 16q22.2 over HYDIN, but without 1q21.1 CNVs. By ddPCR, all had HYDIN CN = 2, but 6 had HYDIN2 duplications, suggesting the presence of 1q21.1 duplications (or other re-arrangements) too small to be reliably detected by microarrays. The exact nature of these variants is not yet determined.
• Twenty-four samples were selected with no microarray evidence of 1q21.1 or 16q.22.2 CNVs (the next two specimens on a plate after each 1q21.1 CNV sample). By ddPCR, all had CN = 2 for HYDIN and HYDIN2.

DISCUSSION
This is the first report of a 1q21.1 duplication carrier with high verbal and low performance IQ. Previous reports noted that developmental and speech delays, intellectual disability and learning disabilities are common in 1q21.1 duplication cases, 2-8 but verbal-performance discrepancies were not mentioned. It is Novel 1q21.1 microduplication phenotype J Xavier et al. Fig. 1 The patient's 1q21.1 duplication CNV and the ddPCR assay strategy for HYDIN2. a Read depth of~40× whole-genome sequencing across 1q21.1 (top panel) and 16q22.2 (bottom panel) for Patient T and HapMap DNA NA12878, plotted using Integrative Genomics Viewer. 24 CNVnator 16 called a heterozygous duplication CNV (copy number = 3, yellow dashed vertical lines) on chr1:146370001-148580000 (2.21 Mb, hg38) for Patient T and no CNV for NA12878. No CNV is observed over HYDIN on 16q22.2 (chr16:70700000-71400000) for either individual. Yaxis: read count from 0-80. b Shown is the region of overlap between the paralogous genes HYDIN2 (1q21.1) and HYDIN (16q22.2, minus strand). Across the~274 kb overlapping region, the two genes are estimated to share 99.4% sequence identity, with 264 Indels and 2049 mismatched bases. 9,25,26 A detailed analysis of the evolution of HYDIN2 demonstrated a more complex set of structural changes than can be shown here. 9 Also shown are the three regions targeted by the ddPCR copy number assays reported here: the two insertions (1 and 6 kbp) that are unique to HYDIN2, and a segment of unique sequence in HYDIN (outside the overlap region) unclear whether this feature is rare, or whether it escaped attention in previous reports which were not focused on comprehensive profiling of individual cases. This case also highlights the presence of motor coordination, articulatory deficits (verbal dyspraxia) and phonological disorders. Phonological deficits were specifically noted by Bernier et al., 8 and Nevado et al. 6 noted an overall phenotypic similarity between 1q21.1 deletion/duplication and 22q11.2 deletion syndromes in which speech disorders are particularly common. 22 Motor difficulties that affect gestural coordination and speech can have a global impact on functioning because of difficulties with socialization and emotional interaction. 23 This case is an example of the great variability of the 1q21.1 phenotype, and it illustrates why comprehensive neuropsychological and behavioral evaluations are needed to guide individualized, multidisciplinary treatment programs.
Whole-genome sequencing did not reveal a molecular basis for the unusual phenotype seen here: the duplication had typical boundaries; no additional neuropsychiatric CNV was detected; and we did not observe notable functional rare variants. We studied HYDIN2 further because it was hypothesized to contribute to cognitive and head size changes in 1q21.1 CNV syndromes. 5 Our patient had normal head size and preserved verbal capacities, thus we wanted to determine definitively whether HYDIN2 was duplicated. We developed two alternative ddPCR copy number assays of HYDIN2 and one of HYDIN; both were concordant with results of dense microarray-based CNV analyses. These assays confirmed HYDIN2 duplication in patient T.
While our ddPCR analyses were in progress, Dougherty et al. published an elegant analysis of the evolution of HYDIN2. 9 To detect structural genomic changes, they evaluated CN using 153 molecular inversion probe (MIP) analyses of single-nucleotide differences between HYDIN and HYDIN2 in 6055 individuals. HYDIN2 was not part of 3 of 25 known 1q21.1 duplications and of 3 of 49 known deletions (consistent with the MGS dataset in which 1 of 7 duplications was distal to HYDIN2). They found that these 6 "atypical" cases had the expected head size changes (microcephaly with deletions, macrocephaly with duplications), suggesting that HYDIN2 may not be involved in these changes. This MIP panel permits more detailed localization of CN change across HYDIN2, but in most cases, HYDIN2 CN can be reliably evaluated with one or both of our ddPCR assays, or with a subset of the above-mentioned MIP assays. Further research will be needed to determine the phenotypic effects of copy number changes in HYDIN2.

DATA AVAILABILITY
The MGS sample was consented for deposition of anonymized clinical information and biomaterials by the National Institute of Mental Health. DNA, clinical information and existing GWAS and CNV data for this sample are available to qualified scientists from the U.S. National Center for Biotechnology Information-dbGAP repository for two data subsets which together, comprise the entire MGS cohort: GAIN (https:// www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id = phs000021.v2.p1) and NON-GAIN (https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi? study_id = phs000167.v1.p1). Genome-wide transcriptome data are also available for a subset https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id = phs000775.v1.p1. Further information is available from the NIMH center for collaborative genomic studies on mental disorders (nimhgenetics.org). Please note that the informed consent from the guardian of Patient T did not include permission to deposit genetic data in a public repository. Investigators conducting related research may contact the present authors to discuss requests for additional information.