Beyond the checkbox: How molecular testing solved my daughter’s "not-quite-autism"

For nearly a decade, my daughter existed in a diagnostic limbo. While clinicians suspected autism spectrum disorder (ASD), years of assessment repeatedly failed to meet the behavioral criteria required by the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5).¹ That struggle ended in the summer of 2021, when a chromosomal microarray (CMA) bypassed the behavioral checklist entirely. The analysis revealed a de novo 466 kb microduplication on 14q11.2 involving the CHD8 and SUPT16H genes. This finding proved she is genotypically autistic, even though she satisfied only 1 minor clinical criterion: deficits in developing and maintaining relationships.

This question addresses a significant clinical dilemma in modern neurodevelopmental medicine: the growing gap between behavioral observation and molecular evidence. When the DSM-5, the gold standard for clinical labeling, conflicts with the objective data of a CMA, it challenges the traditional “phenotype-first” diagnostic model.

As a doctorate-prepared epidemiologist, I am trained to evaluate research literature and leverage emerging technologies to answer complex questions. Yet, despite my professional background and frequent consultations with clinicians, I remained unaware that a CMA could provide the etiological insights that had eluded us for a decade.

The years of clinical purgatory: why the DSM-5 failed us

For a child labeled as “special needs,” the assessment journey is rarely a single event; instead, it becomes a constant loop of reevaluation. This cycle began with her first developmental evaluation at 9 months, followed by formal testing at 2 and 3 years by both a developmental psychologist and a developmental pediatrician. As she transitioned into the education system, her individualized education plan (IEP) necessitated triannual school district reassessments, supplemented by periodic psychological evaluations requested by various psychiatrists. Each new physician expressed skepticism toward the previous “nonautistic” conclusions, yet after their own assessments, they reached the same impasse. While nearly every provider who assessed her believed she was autistic, they were unable to formally diagnose her due to the stringent criteria set by the DSM. This is a frequent clinical challenge in this syndrome, where variable expressivity often leaves patients on the border of a formal diagnosis.² As a parent, navigating the skepticism of each new specialist and the relentless demand to assess and reassess was draining.

Despite the diagnostic frustration, I understood why autism remained a primary consideration. My daughter presented with a complex clinical profile including intellectual disability, executive functioning disorder, attention-deficit/hyperactivity disorder (ADHD), sensory processing disorder, global developmental delay, and delayed speech.³ However, these markers were confounded by her strong social reciprocity and consistent eye contact. Furthermore, she lacked the restricted or repetitive behaviors necessary to reach the threshold for clinical significance under current diagnostic standards.

I felt conflicted by the inability to diagnose her clearly. On one hand, avoiding an ASD diagnosis suggested her prognosis might be less severe, that her challenges were more treatable or temporary. Through a lens of wishful thinking, this ambiguity offered a strange solace. At age 12, following our move to San Diego, California, my daughter’s care transferred to the local children’s hospital. Her new psychiatrist also suspected autism, but was the first to consider a primary genetic etiology for the complex presentation strongly. When she suggested genetic testing, I shared the exhausted history of prior workup, ranging from routine blood tests to an early-childhood muscle biopsy for mitochondrial disorders. Undeterred, she paused and asked a simple but crucial question: "But has she ever had a CMA?”

This diagnostic pivot was driven by a need to reconcile my daughter’s decade of multidisciplinary workups with her consistent failure to meet DSM-5 behavioral thresholds. While behavioral criteria are limited to observable symptoms, a CMA offered a necessary etiological diagnosis by identifying the biological drivers beneath her complex clinical presentation. This approach aligns with current recommendations to utilize genomic testing as a first-tier investigation for idiopathic neurodevelopmental challenges.⁴

She explained that the rationale for targeting a genetic cause was 2-fold: (1) the presence of known high-confidence autism risk genes made the search feasible, and (2) the lack of family history for ASD or atypical neurodevelopmental concerns suggested a de novo mutation.⁵ She colloquially termed this an “act of God,” meaning it was a random, spontaneous genetic change occurring at the time of conception rather than an inherited trait or caused by external factors.

Crucially, modern research has established autism as a highly polygenic and heterogeneous condition.^5,6 This biological complexity underscores the inherent limitations of relying solely on behavioral observation. The psychiatrist noted that even if the CMA did not immediately identify a pathogenic driver, the data would remain in the database for future reanalysis as discoveries emerged. However, future discoveries were unnecessary; the answer that had eluded us for a decade was contained in a 5-page clinical report.

Genotype: the undeniable truth

The verdict: my daughter has approximately 466 kb gain (duplication) on chromosome 14q11.2, spanning the nucleotide sequence 21496133-21962265. Initially, the report formally classified the duplication as a variant of uncertain significance (VUS) per current conservative diagnostic guidelines. However, parental testing revealed a de novo mutation. These findings identified the pathogenic driver for the array of medical conditions we had been navigating, providing the diagnostic clarity that finally unified her complex history into an actionable etiology.

While this microduplication spanned 18 known genes and fell within 5 domains, it was the neurodevelopmental/chromatin domain that provided the etiological anchor for her decade-long clinical limbo.^2,5 Within this domain, chromodomain helicase DNA binding protein 8 (CHD8) and the SPT16 homolog, which facilitates the chromatin transcription complex subunit (SUPT16H), stand out as high-confidence ASD risk genes.

While CHD8 is a highly penetrant risk factor typically associated with a distinct macrocephalic ASD phenotype,⁵ my daughter’s presentation was confounded by a history of symmetrical IUGR and microcephaly. This phenotypic departure underscores the variable expressivity of 14q11.2 microduplication syndrome.³

Functionally, the CHD8 gene acts as a master switch for early brain development.⁵ If CHD8 is the architect, controlling access to the DNA “blueprints” by remodeling chromatin, then SUPT16H helps the crew read those blueprints, facilitating transcription.^2,3 When both genes are duplicated, as seen in 14q11.2 microduplication syndrome, neural network development is fundamentally disrupted.⁸

The genotype-phenotype spectrum

It is estimated that rare or de novo structural variation, primarily in the form of copy number variants (CNVs), is detected in 5% to 10% of ASD families. When broad genomic testing includes single-nucleotide variants (SNVs), rare genetic drivers are identified in up to 15% of families. Despite the identification of over 100 genes associated with ASD, for more than 80% of families, the underlying genetic mechanism remains unknown.⁶

While CHD8 is a well-researched genetic risk factor for ASD, the impact of a CNV involving SUPT16H is less understood. Based on collective observations and detailed patient characterization, a duplication involving the 14q11.2 band, which encompasses both CHD8 and SUPT16H, is described in the literature as 14q11.2 duplication syndrome, with an estimated prevalence of less than 1 per million people.³

My daughter’s microarray report explicitly cited literature showing that an overlapping de novo duplication of CHD8 and SUPT16H is associated with cognitive impairment, speech delay, and ADHD,⁷ the exact constellation of co-morbidities that had confounded her clinical diagnosis for years. To understand why, we must look at the function of these genes:

Because CHD8 regulates broad networks, duplications (extra copies) cause a “ripple effect” across hundreds of downstream genes. This explains why the same genetic error can lead to different psychiatric outcomes depending on which specific networks are disrupted; research shows these CNVs often operate across diagnostic boundaries, conferring risk not only for autism, but for schizophrenia and bipolar disorder as well.⁴

In practice, the 14q11.2 microduplication is defined by variable expressivity, meaning the phenotype can manifest differently across patients. In a 2025 review of 26 patients with 14q11.2 microduplication syndrome, 100% presented with developmental and speech delays, yet only 36% met formal DSM-5 criteria for ASD.³ Additional features identified in this population include intellectual disability (80%), ADHD (40%), and aggressive behavior (53%).³

This data uncovers a significant clinical gap: A child might have a pathogenic mutation in a well-known “autism gene” but show a range of comorbidities, such as global delays, cognitive impairment, and behavioral dysregulation, that do not meet the specific social-communication criteria for a DSM-5 autism diagnosis.

The future of neurodevelopmental diagnostics

The traditional diagnostic process for ASD is built on a “phenotype-first” model: observing behavior, checking criteria, and assigning a label. However, as my daughter’s decade in diagnostic purgatory illustrates, behavior is the end product of a long and winding developmental road—not the starting point.

The value of a genotype-first approach, such as utilizing CMA or whole-exome sequencing (WES) at the first sign of persistent developmental concern, cannot be overstated.⁵ A significant rate of clinical discovery supports this shift: CMA identifies a pathogenic finding in approximately 15% to 20% of patients with unexplained developmental delay or ASD. For children who fall just short of behavioral thresholds, this provides a high-probability path to identifying the underlying biological cause.⁶ Consequently, genomic testing becomes an essential diagnostic tool for children who, like my daughter, present with a complex constellation of comorbidities but fail to meet formal behavioral criteria.

• Ending the “assessment loop”: For families, the “constant loop of reevaluation” is a source of profound trauma and exhaustion. A genomic diagnosis provides an “anchor” that prevents the drift from specialist to specialist.
• Bypassing compensatory masks: Many children, particularly girls, develop “social masking” or have protective factors, such as my daughter’s strong verbal reciprocity, that allow them to “pass” behavioral tests while their underlying neurology remains profoundly taxed. Genetics does not care how well you can make eye contact; it sees the architecture beneath the effort.
• Validation over observation: Modern genetics does not invalidate the DSM-5; it provides a deeper, etiological dimension. If the DSM-5 is the “smoke,” the microarray is the “fire.” The behavioral checklist confirms the presence of an impairment, but the microarray confirms the causation.

Precision intervention: beyond the label

When a primary genetic etiology, such as the 14q11.2 microduplication, is identified, the goal of intervention shifts from chasing symptoms to managing a known biological trajectory.⁵ A genotype-first diagnosis allows for earlier, more targeted interventions. For instance, knowing a child carries a CHD8 duplication allows clinicians to anticipate specific challenges with gastrointestinal issues, sleep disturbances, or executive function deficits that are hallmark “side effects” of this specific genetic driver, even if the child does not meet the DSM-5 threshold for an ASD diagnosis.^3,5

This molecular specificity moves pharmacological management from trial and error toward biologically informed choices. Emerging research on CHD8 duplication models indicates specific dysregulation in the serotonergic and dopaminergic systems, suggesting that the hyperactivity seen in this syndrome may respond differently to medication than standard, idiopathic ADHD.⁸ Furthermore, the diagnosis mandates distinct medical surveillance: Unlike the general population, individuals with this duplication carry a significantly elevated risk for obesity (47%) and epilepsy (27%).³

We can treat the underlying anxiety or the ADHD-like “ripple effects” of the duplication with the confidence that we are addressing the child’s true biological reality, not just a collection of comorbidities.^4,8

Final thought: the gap between code and criteria

My daughter is a perfect example: She is genotypically autistic by carrying a microduplication on a known autism gene, yet she does not meet the behavioral (DSM-5) criteria. Her case highlights a growing disconnect and offers a unique, personal lens into the power of modern diagnostic testing. There is a complex interplay of genetics and environment, directly addressing common misconceptions by explaining how a de novo mutation that is unrelated to family history or external factors can cause autism.

For a decade, we sought a label designed to describe her behavior, yet we eventually identified the biological reality of who she is. Her 14q11.2 microduplication is a primary etiology that predates every behavioral assessment she underwent. Her experience illustrates that while behavioral symptoms define the clinical diagnosis, they are often inextricably linked to the underlying genetic pathways that shape neurodevelopment. This genotypic underpinning offers a critical opportunity: Understanding these genetic drivers can better inform clinical intuition, diagnostic accuracy, and personalized treatment decision-making.

References

1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed, text rev. American Psychiatric Association Publishing; 2022.
2. Smol T, Thuillier C, Boudry-Labis E, et al. Neurodevelopmental phenotype associated with CHD8-SUPT16H duplication. Neurogenetics. 2020;21(1):67-72. doi:10.1007/s10048-019-00599-w
3. Williams JM, Esbit SL, Ho ML, Sidlow R. Neurodevelopmental and behavioral phenotypes in 14q11.2 microduplication syndrome: a case report and literature review. Cureus. 2025;17(8):e90735. doi:10.7759/cureus.90735
4. Thygesen JH, Wolfe K, McQuillin A, et al. Neurodevelopmental risk copy number variants in adults with intellectual disabilities and comorbid psychiatric disorders. Br J Psychiatry. 2018;212(5):287-294. doi:10.1192/bjp.2017.65
5. Bernier R, Golzio C, Xiong B, et al. Disruptive CHD8 mutations define a subtype of autism early in development. Cell. 2014;158(2):263-276. doi:10.1016/j.cell.2014.06.017
6. Eisfeldt J, Higginbotham EJ, Lenner F, et al. Resolving complex duplication variants in autism spectrum disorder using long-read genome sequencing. Genome Res. 2024;34(11):1763-1773. doi:10.1101/gr.279263.124
7. Smyk M, Poluha A, Jaszczuk I, et al. Novel 14q11.2 microduplication including the CHD8 and SUPT16H genes associated with developmental delay. Am J Med Genet A. 2016;170(5):1325-1329. doi:10.1002/ajmg.a.37579
8. Kawamura A, Fujii K, Tamada K, et al. Duplication of the autism-related gene Chd8 leads to behavioral hyperactivity and neurodevelopmental defects in mice. Nat Commun. Published online May 26, 2025. doi:10.1038/s41467-025-59853-5