A microdeletion del(12)(p11.21p11.23) with a cryptic unbalanced translocation t(7;12)(q21.13;q23.1) implicates new candidate loci for intellectual disability and Kallmann syndrome

In an apparently balanced translocation t(7;12)(q22;q24)dn exhibiting both Kallmann syndrome (KS) and intellectual disability (ID), we detected a cryptic heterozygous 4.7 Mb del(12)(p11.21p11.23) unrelated to the translocation breakpoint. This new finding raised the possibility that KS combined with neurological disorder in this patient could be caused by gene(s) within this deletion at 12p11.21–12p11.23 instead of disrupted or dysregulated genes at the genomic breakpoints. Screening of five candidate genes at both breakpoints in 48 KS patients we recruited found no mutation, corroborating our supposition. To substantiate this hypothesis further, we recruited six additional subjects with small CNVs and analyzed eight individuals carrying small CNVs in this region from DECIPHER to dissect 12p11.21–12p11.23. We used multiple complementary approaches including a phenotypic-genotypic comparison of reported cases, a review of knockout animal models recapitulating the human phenotypes, and analyses of reported variants in the interacting genes with corresponding phenotypes. The results identified one potential KS candidate gene (TSPAN11), seven candidate genes for the neurodevelopmental disorder (TM7SF3, STK38L, ARNTL2, ERGIC2, TMTC1, DENND5B, and ETFBKMT), and four candidate genes for KS with ID (INTS13, REP15, PPFIBP1, and FAR2). The high-level expression pattern in the relevant human tissues further suggested the candidacy of these genes. We propose that the dosage alterations of the candidate genes may contribute to sexual and/or cognitive impairment in patients with KS and/or ID. Further identification of point mutations through next generation sequencing will be necessary to confirm their causal roles.

These results indicate that the chromosome 12 breakpoint was within or adjacent to a 3.5 kb interval (chr12: 97,460,688 − 97,464,146 / hg38) between CTD-2542D2 and CTD-2268E11 based on the placement of end sequenced BAC clones on the current genomic sequence map ( Fig. 1B and Fig. 3A). This interval was 874 bp upstream from the 5' end of the RMST locus.
The chromosome 7 breakpoint was mapped within clone CTD-2325L19 (Fig. 2B). Clone RP11-46O13 was telomeric to the breakpoint. Based on the locations of these end-sequenced clones, the chromosome 7 breakpoint is located within an 87 kb interval (data not shown). This interval does not contain any genes and the next nearest gene, ZNF804B, was approximately 11 kb beyond the telomeric boundary of the chromosome 7 breakpoint region de ned by FISH mapping. Cloning of the breakpoint from derivative chromosome 12 The 3.5 kb junction fragment detected on Blot 3 with probes KS-5 and KS-6 from EcoRI digestion from derivative chromosome 12 was ampli ed by two independent suppression PCRs 20 using the primers sets under the condition mentioned in Materials & Methods section. The size of these two PCR products were 0.6 and 0.8 kb respectively, and sequence analysis of these fragments con rmed that they were the junction fragments from the der (12). The genomic breakpoint was located between 97,466,873 and 97,466,877 (hg38) at 12q23.1 (Fig. 4A), within the second intron of RMST (Fig. 1B). As this junction fragment contained chromosome 7 sequences adjacent to the breakpoint, a BAC CTD-2325L19 was identi ed from this sequence by BLAT at Human Genome Browser (hg38), which mapped to 7q21 and was contained in the 3.5 Kb narrowed region by FISH. BAC clone CTD-2325L19 also contained chromosome 7 sequence of this junction fragment and showed hybridization signals of SpectrumGreen on chr7, der(7), and der(12) from FISH (Fig. 2B, right picture). This result indicated that BAC clone CTD-2325L19 contained the 7q21 breakpoint of the subject.
Cloning of the breakpoint from derivative chromosome 7 Because the sequence of chromosome 7 adjacent to the breakpoint of der(12) was now known, the 2.3 kb junction fragment from der(7) was generated by nested PCR using primers proximal to the 7q21.13 breakpoint and ones distal to the 12q23.1 breakpoint using the primers sets under the condition mentioned in Materials & Methods section: Sequence analysis con rmed that this fragment was the junction fragment from der (7). The genomic breakpoint was located between 88,722,752 and 88,722,753 (hg38) at 7q21.13 (Fig. 4A). Two junction fragment sequences were evaluated to determine if there had been any gain or loss of chromosome material at the site of the translocation. Comparison of the normal chromosome sequence at 7q21.13 and normal chromosome sequence at 12q23.1 along with those from the two junction fragments revealed that there had been an unknown 17 bp insertion (GCAATTGCAATGAATAT) in the der(12) junction fragment and 3 bp CTC deletion from chromosome 12 from the der(7) junction fragment 13 (Fig. 4A).
Identi cation of three candidate genes at 12q23.1 from t(7;12) (q21.13;q23.1) We mapped and sequenced both translocation breakpoints to identify a gene that underlies KS. Because two balanced translocations and one deletion 12q24 involving hypogonadism were published 3-5 , sequences upstream and downstream of the chromosome 12 breakpoint were analyzed for the existence of a candidate gene, which resulted in the identi cation of RMST. By comparison of the genomic sequence, it was determined that the translocation directly disrupted RMST such that the 12q23 breakpoint was located within intron 2, downstream of the second exon of this gene (Fig. 1B) 13 . RMST encodes a long noncoding RNA speci cally expressed in the developing brain.
Identi cation of two candidate genes at 7q21.13 from t(7;12) (q21.13;q23.1) In a de novo balanced translocation with associated phenotype, the culprit gene is often located at one of two breakpoints 7,8,10,11 . We have looked at genes located at chromosome 7 breakpoint too. The closest gene to the breakpoint at 7q21. 13 Fig. 2A and Table 2). Validation of putative candidate genes using tissue-speci c RT-qPCR To investigate the functional signi cance of 11 positional candidate genes among 12 in phenotype-relevant tissues, RT-qPCR transcript levels were measured in ve distinct human tissues (brain, fetal brain, muscles, ovary, and testis). The spatiotemporal regulation of gene expression causes varying expression patterns, which also depend on a number of variables, including the RNA isolation process and detection techniques. We used commercially available human RNA samples to measure the expression patterns through RT-qPCR experiments so that we would have a reference of the expression of the genes of our interest. This was done because different expression patterns were found in several publicly available resources, including the GTEx Portal (https://gtexportal.org/home/) and NCBI (https://www.ncbi.nlm.nih.gov/).
Our KS-candidate gene TSPAN11 showed high expression in testis and ovary, whereas ve NDD-candidate genes (TM7SF3, ARNTL2, ERGIC2, TMTC1, and DENND5B) among 7 showed good expression levels in brain and fetal brain. Our four candidate genes for KS + NDD-INTS13, PPFIBP1, REP15, and FAR2 showed expression in ovary and testis along with adult and fetal brain tissues (Fig. 4B). The expression of these genes in tissue samples relevant to the disease-causing organs suggests that they might underlie the clinical phenotype when mutated. The NDD candidate gene ETFBKMT was not included in this experiment, because it is identi ed as an additional ID candidate gene after re-interrogation of all 29 genes at 12p11.21-12p11.23.  4 . In the second case, an apparently balanced translocation t(7;12)(q22;q24) in a male with KS and intellectual disability was published in 1990 2 . Another case reported a balanced chromosomal translocation between the distal q arms of chromosome 4 and 12, t(4;12)(q25;q24.2), in a Turkish male patient in 1994, who showed IHH and a lack of secondary sexual characteristics yet had a normal sense of smell 3 . After the three aforementioned balanced translocations with hypogonadism were published, a de novo interstitial deletion del(12)(q24.3q24.33) was described in a male with ambiguous genitalia and development delay in 1999 5 .
We postulated that all four reported chromosomal rearrangements involving the 12q24 region could be explained by the haploinsu ciency of a gene responsible for KS or IHH. The positional cloning of balanced translocation patients was successful to identify KS genes at the translocation breakpoints 10,11 . To investigate if the translocation might contribute to the phenotypes in the male Subject 1, we mapped and sequenced both translocation breakpoints to clone genes that may be involved in KS seen in this individual with t(7;12)(q22;q24) 13 .
Physical mapping of the translocation breakpoint by FISH and Southern blot hybridization led to the identi cation of ve genes at both breakpoints. ZNF804B mapped 37 kb distal from the breakpoint is the closest gene to the breakpoint at 7q21.13 9 and STEAP4 mapped 416 kb proximal to the breakpoint. The breakpoint at 12q23.1 directly disrupts the non-coding RNA RMST, whereas NEDD1 and PAFAH1B2P2 map closest to the breakpoint on the proximal and distal sides, respectively. Based on the molecular analysis results, the locations of the cytogenetic bands on both chromosomes 7 and 12 of the apparently balanced chromosome translocation have been precisely revised as t(7;12)(q21.13;q23.1) 9,13,14 . We assumed that the causative gene for KS is located on chromosome 12 due to two previously reported chromosomal rearrangements displaying overlapping phenotypes of KS and screened for mutations in three-selected candidate genes (RMST, NEDD1, PAFAH1B2P2) based on their proximity to the 12q23.1 breakpoint. RMST is disrupted in intron 2, and mutation screening of this gene was performed in 48 KS subjects negative to ANOS1 and FGFR1. We identi ed a heterozygous nucleotide change in RMST (heterozygous change of C/C to C/T at the 214th nucleotide in exon 10 of RMST NR_152618.1) in a KS male alongside his mother with anosmia. The nucleotide change turned out to be a polymorphism, as two healthy sisters of the patient shared the same nucleotide change.
Additionally, the two patients in this family were found to have a mutation of FGFR1 (c.821G > A, p.E274G) 26 , suggesting this gene as the cause of the phenotype in this family. PAFAH1B2P2 and NEDD1, mapping 248 kb downstream and 513 kb upstream from the breakpoint, respectively, were also screened in the same collection of 48 KS patients we recruited, with no evidence of mutations.
Two additional genes were mapped in the breakpoint region at 7q21.13. ZNF804B, which has an unknown function, is located 37 kb downstream from the breakpoint; hence, it is the closest gene to the breakpoint 9 . ZNF804B is a member of the zinc nger protein family, which is composed of four exons. ZNF804B has not previously been described as a disease-related gene with substantial evidence. There is limited information in the literature regarding its full biological functions, except it was found to be one of the candidate genes associated with autism spectrum disorder and neurodevelopmental disorders 27,28 . This suggests that ZNF804B might be dysregulated by position effect and could be the cause of intellectual disability of Subject 1. In addition, STEAP4 at 7q21.12 is located 416 kb upstream from the breakpoint. Functioning as metalloreductase, STEAP4 resides in the Golgi apparatus and may be involved in adipocyte development and metabolism. Mutation screening of both genes in the same cohort of 48 patients we recruited did not detect any putative mutations, though polymorphisms were identi ed. Additionally, we screened for mutations in ANOS1 and FGFR1 in the t(7;12) translocation subject. No potential disease-causing mutations were found (data not shown). We cannot exclude the possibility of mutation of other causative gene(s) for KS in this subject.
Six percent of carriers of balanced translocations manifest abnormal phenotypes 29 due to the disruption of the genes at the breakpoints or dysregulation (i.e., position effect) resulting in reduced expression of the genes by the separation from its cis regulatory elements 30 . Nonetheless, 40% of the patients with apparent balanced translocations were reported to have at least one deletion at one of the breakpoints or genomic regions elsewhere, suggesting that deletions might be common in apparently balanced chromosome rearrangements 16 .
As we did not nd any mutations in ve candidate KS genes at or near both breakpoints from the apparent balanced translocation, we performed aCGH analysis, which remarkably unmasked a cryptic 4.7 Mb submicroscopic microdeletion at 12p11.21-12p11. 23. This cryptic microdeletion was not detected in previous studies of the breakpoints of Subject 1 (DGAP032) on the molecular level 9,13,14 .
There has been a report of chromosomal variation including CNVs in lymphoblastoid cell lines (LCLs). In their study, Shirley et al found more CNVs per sample on average for LCLs than for PBMCs, but the differences were statistically not signi cant.
Furthermore, chromosomal regions associated with a signi cantly different number of CNVs did not include chromosome 12 31 .
Thus, it is unlikely that del(12)(p11.21p11.23) is a culture induced artifact, although we cannot exclude this possibility entirely. Unfortunately, the de novo status of this microdeletion could not be determined due to the unavailability of parental samples.
In a patient with two different concomitant genomic rearrangements such as an unbalanced translocation and a simultaneous translocation-unrelated microdeletion or microduplication, the culprit gene was found to be located at one translocation genomic breakpoint sometimes albeit rare. For instance, in a female subject affected with autism and ID with t(14;21)(q21. This microdeletion encompassing 29 genes is likely to harbor the disease genes involved in the common set of neurodevelopmental phenotypes shared with our seven studied CNV patients and eight unpublished CNV cases from the DECIPHER database ( Fig. 2A, Tables 1 and 3).
Among 15 heterozygous CNVs at 12p11.2, duplications account for ten cases, whereas deletions represent the remaining ve cases ( Fig. 2A, Tables 1 and 3). At least one neurodevelopmental phenotype was seen in each case. Some of the CNVs were inherited from one parent with an unknown phenotype, whereas the inheritance of the remainders is unknown (Tables 1 and 3).
This suggests that some genes in the CNVs located at 12p11.2 may have incomplete penetrance or epigenetic imprinting if a carrier parent is asymptomatic. We interpret the pathogenicity of deletion and duplication cases similarly since they both exhibit the similar neurological condition. This is because any CNV might interfere with the strict stochiometric control of genes on a protein expression level 36 .
Based on in silico comparative genomic analysis at 12p11.21-12p11. 23 73 . The mutations of the former cause Lenz microphthalmia, an Xlinked syndromic intellectual disability 74 ; however, the variants of the latter were found in individuals with ASD 45 , developmental delay 28 , schizophrenia 45,46 , and Tourette syndrome 75 .
At 12p11.21, we identi ed two additional ID candidate genes. One ID candidate gene is DENND5B (DENN Domain Containing 5B, MIM 617279), a guanine nucleotide exchange factor (GEF) mediating the activation of small GTPases. With many downstream targets, they function as molecular switches in intracellular signaling pathways. There are some known GEFs involved in NDDs.
TUBB2A is associated with seizures, ID and development delay 92 , while TUBB4A mutations cause leukoencephalopathy hypomyelination with atrophy of the basal ganglia and cerebellum 93 . DARS2 is genetically linked to leukoencephalopathy with brain stem and spinal cord involvement 94,95 124 . We excluded both genes due to the phenotype unrelated to KS and ID.
We propose that, despite not being encompassed by small CNVs we used in comparative genomic mapping, the expression levels of the two ID candidate genes DENND5B, and ETFBKMT could be altered due to position effect 125,126 , providing a likely explanation for NDD phenotypes such as dystonia, global developmental delay, growth delay, motor delay, etc., observed in one DECIPHER proband DCP288321 ( Fig. 2A and Table 3).
We also con rmed the high expression of our candidate genes in ve different human tissues (i.e., brain, fetal brain, muscle, ovary, and testis) relevant to the phenotype of KS and NDD, to substantiate their pathogenicity (Fig. 4B).
During the composition of this manuscript, genome sequencing was performed in this same patient to map the translocation breakpoint, with the conclusion that deletion of RMST was implicated as a cause of KS through loss of function by the erroneous assumption that this chromosome translocation is balanced 14 . RMST physically interacts with SOX2 127 , a transcription factor known to regulate neural fate, and aids in the binding of SOX2 to the promoter of target genes important in neurogenesis 127 . SOX2 (SRY-box transcription factor 2, MIM 184429) is a known disease gene for hypogonadotropic hypogonadism and combined pituitary hormone de ciency 1 . RMST has also been associated with rhabdomyosarcoma and melanoma 128 .
This ostensible pathogenicity of RMST in KS remains to be seen, because this subject has an unbalanced translocation accompanied by an additional 4.7 Mb microdeletion that we identi ed. Moreover, we did not nd any mutations of this gene in the 48 KS patients we recruited. Our case underscores the necessity and signi cance of aCGH or sequencing analysis in individuals with disease-associated apparently balanced translocations to rule out cryptic microdeletions. At the same time, this study highlights the bene t of the integrated usage of karyotype analysis, aCGH and sequencing for an informed approach to phenotypic assessment.
In summary, we found that an apparently balanced translocation t(7;12)(q22;q24) 2,13,14 is actually unbalanced and the 4. 7 Mb cryptic deletion at 12p11.21-12p11.23 we identi ed is likely to explain the phenotype of KS and ID in the subject carrying these two unrelated chromosomal rearrangements. In silico comparative genomic mapping with additional 14 CNVs in this genomic region identi ed one potential KS candidate gene (TSPAN11), seven candidate genes for neurodevelopmental disorder (TM7SF3, STK38L, ARNTL2, ERGIC2, TMTC1, DENND5B, and ETFBKMT) and four candidate genes for KS with ID (INTS13, REP15, PPFIBP1, and FAR2). The candidacy of these genes was further supported by the high-level expression pattern in the relevant human tissues. We propose that some dosage-sensitive genes, either increased or decreased, in this genomic region might contribute to the sexual and/or cognitive impairment in the patients with KS and/or ID. Among our candidate genes, the probabilities of dosage sensitivity, such as pHaplo and pTriplo of at least three genes, are high (STK38L: pHaplo/pTriplo 0.77/0.94, DENND5B: 0.96/0.96, PPFIBP1: 0.8/0.7) 129 , suggesting that both its increase or decrease can result in a related deleterious phenotype. RT-qPCR or western blot of candidate genes encompassed in both a deletion and duplication will help generate substantiating evidence for this hypothesis.
It is well known that heterogeneous neurodevelopmental phenotypes are caused by mutations in the same gene 130 . Given that our seven CNV subjects show diverse neurodevelopmental phenotypes including intellectual disability, autism, and epilepsy, our candidate genes at 12p11.21-12p11.23 will offer an opportunity to identify NDD disease genes from NGS databases containing a myriad of autosomal dominant or de novo VUSs. A detailed description of this DGAP032 subject with a balanced de novo reciprocal translocation, 46,XY,t(7;12)(q22;q24)dn was previously published in 1990 2 . In brief, the patient was a 44-year-old Chippewa/French man in 1990 with hypogonadotropic hypogonadism, based upon low levels of FSH, LH, and testosterone, along with sparse pubic hair, small testes (< 1 cm), de ciency of olfaction, skeletal and cranial anomalies, and ID. He was referred to a hospital at the age of 22 years due to delayed sexual development. During this evaluation, he was noted to have normal 17-hydroxycorticosteroids and abnormally low 17ketosteroids and gonadotropin levels. The epiphyseal centers of most long bones and the spine were not yet closed. The bone age of the hand was in the "neighborhood" of 12 years, and the metacarpals appeared shortened and clubbed at their distal ends, especially the 4th right metacarpal, indicating brachydactyly. In addition, a sharply outlined foramen in the occipital bone near the internal occipital protuberance was noted. In 1984, lymphocyte chromosome studies demonstrated an apparently reciprocal translocation, t(7;12)(q22;q24), which was revised as t(7;12)(q21.13;q23.1)dn based on our molecular analysis result ( Fig. 1A) 13 . Clinical signs of KS were not seen in his ve full sisters, as well as one full brother, two half-brothers, or one halfsister.

Human Subjects
Subject 2-500 kb dup(12)(p11.23) Subject 2 (50943) is a 36-year-old man with a history of ID, developmental delay, autism, and dyslexia. The delivery took place with a birth weight of 3500 g. His IQ, assessed at 48 months, was between 30-50. He sat independently at around 11 months, walked at 27 months, was toilet-trained at 2 years, and spoke his rst words at 30 months. He was also noted to have learning disability, language delay, and speech delay. At 29 years old, he was noted to speak often loudly, and demanding frequently with incomprehensible associations. His tantrums are verbal but seldom physical and sometimes very incriminating. He is extremely restless with no tics or stuttering, constantly repeats words and parrots "yes." He speaks very loudly. Although no apparent physical abnormalities were observed, he was noted to have hypertelorism and a slightly thicker lower lip. He did not have seizures, bone anomalies, facial dysmorphism, or shortened ngers or toes (Fig. 5). However, a brain MRI displayed an arachnoid cyst. aCGH analysis carried out on genomic DNA revealed a de novo 500  Subject 3 (31606) is a 10-year-old female with a history of developmental delay, speech delay, expressive language delays, and ADHD. At the age of 15 months, she managed to walk; however, she had occupational therapy because she would not crawl, and she also required social/emotional therapy. She was enrolled in early childhood education and speech therapy and attended a mainstream school at 4.5 years old. She had an aversion to textured foods as a baby requiring feeding therapy. At 3 days of age, she had chronic diarrhea and did not grow or gain weight well. She was hospitalized four times from 52 days until 5 months for failure to thrive and chronic diarrhea. She was hospitalized a third time for 18 days with no conclusive diagnosis despite testing. She was chronically dehydrated because her body could not absorb necessary nutrients as uids passed through her quickly. A second gastroenterology opinion was sought; however, no additional tests were conducted. She also had congenital sucroseisomaltase de ciency (CSID Subject 4 (022821) is a four-year-old white male with a history of learning disability, dyslexia, hypotonia, language delays, and delayed speech. Born at full-term, via normal spontaneous delivery, he crawled at 7 months, sat alone at 9 months, and walked unassisted at 2 years of age. He required time to be able to sit alone because he had no strength in his abdomen. He took his rst steps at 26 months, and since 30 months, has been receiving occupational therapy (OT). Evaluated by a neurologist at 15 months, he was diagnosed as having microcephaly (OFC 43 cm, approximately − 3SD for age), short stature, developmental delay, and impaired motor skills. He also presents with syndactyly and tapering ngers. At age 2 years, some visual asymmetry in visual evoked potentials was detected. CMV and other intrauterine infections in the mother were excluded. He was always attentive to his surroundings, but with no attempt to communicate. At 3 years and 10 months, it was discovered that he had moderate hearing loss in the left ear and mild hearing loss in the right ear (Fig. 5). An EEG done at age 23 months was normal.
Chromosomal analysis was normal, but a SurePrint-Ga Human Genome Kit Agilent aCGH (4x180K) revealed a 1. Subject 5-215 kb dup(12)(p11.23) Subject 5 (DCP295472) is 12 years and 5-month-old Caucasian male with a history of autism spectrum disorder and learning disability. He was born at full-term by spontaneous vaginal delivery and weighed 3.810 kg (76th centile), length of 50 cm (38th centile), and had a head circumference of 34 cm (22nd centile). He suffered from repetitive ear infections in childhood and underwent a surgery for ear tubes. He walked at 24 months and had a language delay associated with a global developmental delay and impaired motor skills. Neurobehavioral concerns began at 2 years when he displayed repetitive movements and stereotypy. He suffered from di culties regulating emotions and had few facial expressions. Sleeping disorders occurred with early awakening at 3-4 am treated by melatonin. He expressed food selectivity. At 10 years of age, he weighed 30.8 kg (median) with a height of 135 cm (median) and head circumference of 53 cm (median). He had right eyelid ptosis. While an ophthalmic surgery was planned, the subject never came to the anesthesiologist appointment. Dysmorphic features included bilateral downward palpebral ssures, right eyelid ptosis, frontal hair spike and lower lip eversion. He also presented with clinodactyly of the 5th nger in the right hand (Fig. 5 Subject 6 (DCP370033) is an 11-year-old Caucasian girl with a history of developmental delay, speech delay, learning di culties, and ADHD. She is the rst child of a healthy and non-consanguineous couple. The pregnancy was uncomplicated. She was born full term by spontaneous vaginal delivery and weighed 3 kg (99th centile). She has two younger sisters, one of whom has a unilateral third nger brachydactyly, according to her mother. The patient sat independently at approximately 7 months, walked at 22 months, and spoke her rst words late. At age 5 years, speech delay with articulation problems became apparent.
Psychological tests revealed a clear discrepancy between performance and verbal capacities. The mother described behavioral challenges including temper tantrums and a short attention span. She is presently in the 3rd grade, has mild ID and learning di culties, shows a poor attention span, and is easily distracted. She is in a special education program, with curricular adaptations. She is seen in a neuropediatric clinic for recurrent headaches that improved after withdrawing methylphenidate treatment and is in an endocrinology clinic for obesity. She has normal stature, normal head circumference and no dysmorphic features. She presents with a large thumb, shortening of the IV and V metacarpals and of the IV and V metatarsals, tapering ngers without clinodactyly or syndactyly. She has short toes and a short 4th metatarsal bone (Fig. 5). Her father with the same microdeletion had some learning di culties and with the same abnormal feet, but he nished the 12th grade as did her uncles, who were not tested for aCGH. The father has at least two brothers with more severe learning di culties, but they are able to live independently. aCGH (CGX-HD 180K (PerkinElmer®)) performed shows a paternally inherited 2. Subject 7 (DCP293962) is a 48-year-old Caucasian male with a history of developmental delay, dyslexia and ID with poor academic performance. He was born full term by spontaneous vaginal delivery with an average birth weight. His neonatal period was unremarkable. He had normal gross and ne motor milestones. He was an average child regarding his social interactions. He was diagnosed with dyslexia in early childhood and has speech delay. He could not read and can barely write his name. He left school in grade 8 and has been on a disability pension since then because he had di culties holding a job. He does not have any disruptive or aggressive behavior. He has recently married at the age of 42. At this age, his height was 179 cm, weight was 95 kg and his head circumference was 59.2 cm. He was dysmorphic showing upturned nostrils and a high nasal bridge. He presents clinodactyly and tapering ngers (Fig. 5). The subject's examination was unremarkable. His fragile X and urine metabolic screen were normal. His aCGH showed a paternally derived microduplication: Eight Decipher CNV patients in Fig. 2A Brief clinical information, genomic coordinates, and inheritance patterns of eight Decipher CNV patients used in silico comparative genomic mapping in Fig. 2A are described in Table 3.

Fluorescence in situ hybridization analysis
A lymphoblastoid cell line (GM10565) from Subject 1, designated DGAP032 in the Developmental Genome Anatomy Project, was obtained from the NIGMS Human Genetic Cell Repository at the Coriell Institute for Medical Research (www.coriell.org) 12 . The karyotype, 46,XY,t(7;12)(q22;q24)dn, was recon rmed prior to re nement of the breakpoint by FISH. Assignment of chromosome breakpoint locations to chromosomal bands was determined by GTG-banding. To identify genes potentially disrupted in the subject, translocation breakpoints were mapped using FISH. Maps from the National Center for BiotechnoIogy Information (http://www.ncbi.nlm.nih.gov/genome/guide/human/) 131  corresponding to relative locations on the UCSC map from chromosomes 7 and 12 were used as FISH probes on metaphase chromosome spreads from an Epstein-Barr virus-transformed lymphoblast cell line generated from the subject`s peripheral blood. Metaphase spreads were prepared according to a standard cytogenetic protocol. Human BAC clones were obtained from the RP11 (Children's Hospital of Oakland Research Institute) and the CIT pool D (Research Genetics) libraries. BAC DNA was puri ed by alkaline lysis and isopropanol precipitation. After puri cation, BAC DNA was directly labeled by nick-translation with either SpectrumOrange or SpectrumGreen labeled nucleotides (Vysis) and used in single-or two-color FISH experiments. Slides were counterstained with 4',6'-diamidino-2-phenylindole hydrochloride (DAPI). Representative metaphase images were recorded using the CytoVision image analysis system (Applied Imaging) database.
Using the relative STS positions on the UCSC map, BAC clones were chosen to cross the relevant regions on chromosomes 7 and 12. FISH analysis of each clone was then used to identify clones that mapped proximal or distal to each chromosome breakpoint. By this way, physical maps of chromosomes 7q21 and 12q24 were constructed, and the breakpoint regions narrowed and de ned 13 .

Southern blot analysis
Southern blot analysis of patient lymphoblast genomic DNA with seven probes (from KS-1 to KS-7) to search for altered restriction fragments was carried out using standard protocols (Fig. 3B). For each lane, 10 µg of genomic DNA from the patient and control were digested with an appropriate restriction enzyme. Fragments were separated on a 1.0% agarose gel and transferred to Hybond-N membrane (Amersham, Arlington Heights, Illinois, USA). Filters were ultraviolet cross linked, baked at 80°C, and hybridized with probes labelled with 32P-dCTP by random priming. Hybridization of labelled fragments was done in the presence of excess herring sperm competitor DNA, and hybridized membranes were washed at 60°C with 0.15 M NaCl/0.015 M sodium citrate/0.1% sodium dodecyl sulphate (SDS) for 30 minutes. Autoradiography took place for 16 hours at − 70°C using two intensifying screens. Seven hybridization probes were ampli ed by the primer sets mentioned in Supplementary table 1. After the breakpoint region was apparently narrowed to 3.5 kb between CTD-2268E11 and CTD-2542D2 at chromosome 12 in band q23 by FISH, the rst four genomic probes, KS-1, KS-2, KS-3, and KS-4, within this region were ampli ed from the breakpoint spanning BAC clone RP11-492N15 under the following conditions (Fig. 1B, Fig. 2B, and Fig. 3A):: initial denaturation at 94°C for 2 min, followed by 30 cycles at 94°C for 30 sec, 58°C for 30 sec, 72°C for 45 sec (KS-1, KS-2, and KS-3) or 3 min for 30 sec (KS-4), and extension at 72°C for 5 min after the last cycle. Two genomic probes KS-5 and KS-6 were ampli ed from this region (Fig. 3A) using BAC RP11-492N15 under the following conditions: initial denaturation at 94°C for 2 min, followed by 30 cycles at 94°C for 30 sec, 63°C for 30 sec, 72°C for 50 sec, and extension at 72°C for 5 min after the last cycle. The probe KS-7 was ampli ed within that putative breakpoint region using BAC RP11-492N15 under the following conditions: initial denaturation at 94°C for 2 min, followed by 30 cycles at 94°C for 30 sec, 58°C 30 sec, 72°C for 40 sec, and extension at 72°C for 5 min after the last cycle. The list of primers used for the ampli cation of probes for Southern hybridization is presented in Supplementary Table 1. Suppression PCR and Nested PCR The 3.5 kb junction fragment from der (12) was ampli ed by suppression PCR using the following primer sets and the conditions. Primers anking the 1.5 kb narrowed breakpoint region from the 12q23.1 were used with adaptor-based primers: PCR1): AP1-A 5'CCTAATACGACTCACTATAGG3' + AC007351-54209rev 5'GTGAATGGTGGATAGTGCTC3'; AP2-A 5'CTATAGGGCTCGAGCGGC3' + AC007351-54176rev 5'GATTAAATTCACTCTCTGAAGAA3'. PCR2): AP1-A 5'CCTAATACGACTCACTATAGG3' + AC007351-54176rev 5' GATTAAATTCACTCTCTGAAGAA3'; AP2-A 5'CTATAGGGCTCGAGCGGC3' + AC007351-54026rev 5'CTAGCTTACAATTTTCTGGTGA3'. Initial denaturation was at 94°C for 2 min, followed by 30 cycles at 94°C for 30 sec, 57°C for 30 sec, 72°C for 1 min 30 sec, and extension at 72°C for 5 min after the last cycle. The 2.3 kb junction fragment from der(7) was ampli ed by nested PCR using the following primer sets and the conditions. Initial denaturation at 94°C for 2 min, followed by 30 cycles at 94°C for 30 sec, 57°C for 30 sec, 72°C for 2 min 30 sec, and extension at 72°C for 5 min after the last cycle. PCR1) 13 : 5'CCATTGGCTTTAAGTGTATAGT3'+ 5'CTTGTGTGTACATCTCCTGAA3'; PCR2): 5'CAACAGACATCTGCATTTACTT3'+5'GAAGATAGCTATAACAACAGC3'.
Mutation screening of ve genes at the breakpoints of t(7;12) (q21.13;q23.1) We screened for mutations in ve genes -RMST, NEDD1, PAFAH1B2P2, ZNF804B, and STEAP4in 48 KS patients including our translocation subject, who was also screened for FGFR1 (Fibroblast Growth Factor Receptor 1, MIM 136350) and ANOS1 (Anosmin 1, MIM 300836) to exclude the possibility of mutations in these two well-known genes with high prevalence for KS. A combination of single strand conformation polymorphism analysis (SSCP) and direct sequencing of RMST and ZNF804B were performed in mutation screening. The 46 primer sets were designed to cover all exons and anking intronic regions of two predicted mRNAs. The size of amplicon is adjusted to less than 350 bp for SSCP, and a few parts of mRNAs were applied on PCR and direct sequencing to check for a mutation. PCR was carried out with 10 ng of genomic DNA of the subject's sample or normal control in 20 ul of reaction (primer sequences and ampli cation conditions are available on request). Then PCR products were electrophoresed on precast gels of ExcelGel DNA Analysis Kit (Amersham Biosciences) according to the manufacturer's instructions. Sequentially, the DNA gel was stained by silver stain according to the manufacturer's instructions (DNA silver staining kit; Amersham Biosciences) to visualize and permanently stain the discrete DNA bands. When aberrant band patterns were recognized on the samples compared with normal controls, PCR products that have the aberrant band on the gel were sequenced with ABI Prism 377 sequencer (Applied Biosystems, Foster City, Calif.). Sequences were aligned and compared with sequences of predicted mRNAs to con rm the mutation. For seven genes, all coding regions and exon-intron boundaries were directly ampli ed and sequenced. NCBI reference mRNA sequences used for screening were NR_152618.1 (RMST), NM_001135175.1 (NEDD1), NR_077240.1 (PAFAH1B2P2), NM_181646.5 (ZNF804B), NM_024636.4 (STEAP4), NM_001174067.1 (FGFR1), and NM_000216.4 (ANOS1).

aCGH (Array Comparative Genomic Hybridization)
DNA extracted from the cell line was compared to a reference sample for standard two-color aCGH. Reference DNA was purchased from Promega (Madison, WI, USA). Test samples were labeled using Cy5 and reference DNA was labeled using Cy3. Agilent 244K human genome oligonucleotide aCGH (G4411B) was used for aCGH analysis following the manufacturer's instructions (Oligonucleotide Array-Based CGH for Genomic DNA Analysis protocol version 3 (Agilent Technologies, Palo Alto, CA, USA). Images were captured using an Agilent scanner and quanti ed using Feature Extraction software v9.0 (Agilent Technologies, Palo Alto, CA, USA). CGH analytics software v3.4 (Agilent Technologies, Palo Alto, CA, USA) was subsequently used for data normalization, quality evaluation, and data visualization. Copy number aberration was indicated using the ADM-2 (Aberration Detection Method 2) algorithm. Probe positions were mapped to GRCh38.
In Silico Comparative CNV mapping The phenotypes from our seven CNV patients (Subjects 1-7), DGAP032, 50943, 31606, 022821, 295472, 370033 and 293962 (Table 1) were compared with eight unpublished CNV cases from the DECIPHER database. Genomic coordinates from these cases were converted to hg38 before the comparison was carried out (Table 3). Three factors were considered for choosing candidate genes -(1) sporadic genetic variants reported in humans with matching phenotypes, (2) knockdown or knockout animal models recapitulated human phenotypes, (3) their interacting proteins were investigated, and the genetic variants of corresponding genes reported in human patients with a similar phenotype (Table 2). Literature was also reviewed as well as several databases including Human Gene Mutation Database HGMD Professional (2022.2) (https://my.qiagendigitalinsights.com/bbp/view/hgmd/pro/start.php), MGI (6.21) (Mouse Genome Informatics) (http://www.informatics.jax.org/), BioGrid (4.4.212) (Database of Protein, Chemical, and Genetic Interactions, thebiogrid.org), and VarElect (https://varelect.genecards.org/). Figure 1 (A) Ideogram illustrating the revised t(7;12)(q21.13;q23.1)dn karyotype in subject 1, DGAP032. After breakage of two chromosomes, the reciprocal exchange of chromosome segments between chromosomes 7 and 12 has taken place, generating two derivative chromosomes in the subject with two horizontal gray bars at the breakpoint positions. On chromosome 12, the deleted cryptic segment at 12p11.21-12p11.23 identi ed was depicted as a horizontal yellow bar. (B) Physical mapping of the 12q23 translocation breakpoint of DGAP032 by FISH and Southern blot hybridization. Diagram shows breakpoint re ned by FISH and Southern blot. For FISH, two BAC clones RP11-492N15 and CTD-2235H23 spanning the breakpoint, which is represented as a dashed vertical red line, were identi ed, and shown as red bars. The breakpoint was further narrowed to 3.5 kb between CTD-2268E11 and CTD-2542D2 shown as black bars. Southern blot analysis using Blot 4 with the probe KS-7 identi ed aberrant fragments of subject DNA digested with ve different restriction enzymes (DraI, BbVI, MboI, PvuII, and HaeIII, Figure 3B). The breakpoint was re ned to 492 bp between the centromeric end of HaeIII and the telomeric end of DraI, which was then isolated with suppression PCR and sequenced. The breakpoint at 12q23.1 is located in intron 2 of RMST (NR_152618.1). Figure 2 (A) Cryptic 4.7 Mb heterozygous deletion encompassing 29 genes located from 12p11.21 to 12p11.23. Eight heterozygous CNV cases from the DECIPHER database are denoted by DCP along with six heterozygous CNVs (subjects 2-7) we recruited. These 14 CNVs are encompassed in subject 1 to help narrow down the candidate gene region by in silico comparative genomic mapping. Red bars represent deletions, whereas blue bars represent duplications. One gene in green is a candidate for Kallmann syndrome, whereas seven genes in brown are NDD candidates. Four genes in purple are chosen as candidates for KS combined with ID. Arrow indicates transcriptional direction of each gene. (B) FISH with two BAC clones spanning the breakpoints at 12q23.1 and 7q21.13, respectively. BAC clone RP11-492N15 shows normal signals on normal chromosome 12 and split signals on both derivative chromosomes 7 and 12. (2). BAC clone CTD-2325L19 with normal signals on normal chromosome 7 and split signals on both derivative chromosomes 7 and 12.  (A) Sequences of the junction fragment composed of two different chromosomes at the translocation breakpoints. Sequence comparison of the normal chromosomes 7 and 12 with der(7) and der(12) at the breakpoints junctions. Three bp sequence CTC from chr12 is deleted at the junction of der (7), and a 17 bp insertion was found at the junction of the der(12). (B) Transcript levels of INTS13, TM7SF3, STK38L, ARNTL2, PPFIBP1, REP15, FAR2, ERGIC2, TMTC1, TSPAN11, and DENND5B in ve different human tissues (i.e., brain, fetal brain, muscle, ovary and testis) were determined by RT-qPCR. Varying levels of expression of these candidate genes were detected in different tissue samples.   Table 1 shows (a) microcephaly at one year (b) lateral view of right face at 3 years 10 months (c) hypoplasia of the vertical or Thenar palmar exion creases (PFC) on the right palm (d) the vertical PFC looks short and the transverse proximal looks with some tendency to Sidney line on the left palm (e) tapering ngers (f) a minor syndactyly between toes 2 and 3 (g) small toes and a questionable gap between toes 1 and 2 and S5: Subject 5 on Table 1 shows (a) small forehead, right eyelid ptosis, bilateral downward palpebral ssures, frontal hair spike, lower lip eversion (c) tapering ngers and bilateral short 5th ngers with clinodactyly (d) the transverse proximal PFC has tendency to join the transverse distal, a variant of the PFC. S6: Subject 6 on Table 1 shows (a,b) short 4th toes likely due to short 4th metacarpals on both feet (c,d) dorsal view of tapering ngers with short 4th and 5th ngers in both hands (e) the transverse distal palmar exion crease is rather short and doesn't start at the ulnar margin on both palms (f) dorsal view of both hands showing bilateral short 4th ngers and short 5th ngers especially at right S7: Subject 7 on Table 1 shows (a) essentially no dysmorphism (b,c) microcephaly (d) tapering ngers (e,f) clinodactyly of the 5th toes and a gap between toes 1 and 2 in the left foot.