Autism spectrum disorder is a heterogeneous neurodevelopmental condition defined by differences in social communication and by restricted, repetitive patterns of behavior or interests. It affects roughly one in 100 children worldwide, yet its prevalence in African populations remains inconsistently documented. Decades of genomic research have established that autism is strongly genetic, with heritability estimates between 60 and 90 percent, driven by a combination of common polygenic risk and rare, high-impact variants. Despite this, African populations, which harbor the greatest genetic diversity on Earth, remain markedly underrepresented in sequencing studies and reference databases, leaving variant interpretation in these communities imprecise and diagnosis incomplete.

The Rwandan team recruited 34 children with a clinical diagnosis of autism, established through DSM-5-TR criteria and the Autism Diagnostic Interview-Revised, along with both biological parents for each child. After first-tier Fragile X syndrome screening identified three positive cases, the remaining 31 trios underwent whole-exome sequencing on an Illumina NovaSeq6000 platform. The sequencing achieved a median coverage depth of approximately 200-fold, with more than 97 percent of targeted regions covered at greater than 30-fold depth, ensuring high confidence in variant detection. Analyses were performed under ISO 15189-accredited protocols using clinically validated bioinformatic pipelines, with copy-number variants detected by an in-house tool and all variants classified according to ACMG and ClinGen standards.

The children in the cohort presented with a striking burden of additional neurodevelopmental features. Intellectual disability was present in nearly 97 percent of participants, delayed speech and language development in more than 80 percent, and global developmental delay in over half. Seizures affected roughly a quarter of the children, and dysmorphic features appeared in nearly a quarter. Most cases were classified as severe autism. This clinical complexity reflects the syndromic end of the autism spectrum, where autism occurs alongside intellectual disability, epilepsy, congenital anomalies, or other neurological signs, and it is precisely in such cases that exome sequencing tends to be most diagnostically productive.

The sequencing effort identified 11 rare genomic findings of potential clinical significance in nine patients. Four of these were classified as likely pathogenic, yielding a diagnostic rate of 12.9 percent, a figure that rises to 29 percent when variants of uncertain significance with strong phenotypic concordance are included. That range sits comfortably within the 10 to 30 percent diagnostic yields reported in large trio-based sequencing studies of autism in Europe and North America, suggesting that the genetic underpinnings of autism in Rwandan children are broadly comparable to those described elsewhere, even as the specific variants differ.

Among the likely pathogenic findings were two de novo single-nucleotide variants in well-established autism genes. One child carried a de novo missense variant in GABRB3, a gene encoding a GABA receptor subunit critical for inhibitory neurotransmission; the child’s early-onset seizures, intellectual disability, and language delay align with developmental and epileptic encephalopathy linked to this gene. Another child harbored a de novo splice-region variant in SYNGAP1, a key regulator of synaptic plasticity, in a presentation consistent with the global developmental delay and intellectual disability characteristic of SYNGAP1-related disorders. Both discoveries reinforce the central role of synaptic dysfunction in autism pathogenesis.

The study also uncovered two likely pathogenic copy-number variants. One patient carried a large de novo deletion on chromosome 1p35.3-p35.2 spanning dozens of genes, including dosage-sensitive contributors to neurodevelopment such as PUM1, SDC3, and MECR, associated with autism, hypotonia, seizures, and developmental delay. A second patient carried a deletion within the GNAS locus on chromosome 20q13.32, inherited from an unaffected parent, a finding that highlights the interpretive complexity of inherited variants and the roles of variable penetrance and expressivity in autism genetics.

Beyond the likely pathogenic results, the team identified seven variants of uncertain significance in genes including SHANK3, SYNJ1, KIF15, HUWE1, and EPG5. Several showed notable phenotypic overlap with known disorders: a de novo SHANK3 variant in a child with autism, developmental delay, and craniofacial anomalies echoes the spectrum of Phelan-McDermid syndrome, while compound heterozygous SYNJ1 variants in a child with epilepsy and intellectual disability fit with the gene’s established role in synaptic vesicle recycling. These variants cannot yet be deemed causative, but their concordance with published phenotypes makes them strong candidates for future functional studies and reclassification as evidence accumulates.

The researchers emphasize that the study’s single-site design and modest sample size limit statistical power, and that exome sequencing cannot capture non-coding regulatory variants or all structural variation. More fundamentally, the reliance on reference databases such as gnomAD and ClinVar, which incompletely represent African populations, likely inflated the proportion of variants of uncertain significance observed. Expanding African genomic reference datasets, the authors argue, is essential to improve variant classification, sharpen diagnostic accuracy, and ensure that the benefits of genomic medicine are distributed equitably rather than concentrated in well-studied populations.

Even with these caveats, the study marks a meaningful advance. It demonstrates that trio-based exome sequencing is feasible and clinically useful in Rwanda, capable of shortening the diagnostic odyssey for families, informing medical management, and clarifying recurrence risks for future pregnancies. The path forward, the authors conclude, lies in continued inclusion of ancestrally diverse populations, functional validation through RNA sequencing, splicing assays, and patient-derived neuronal models, and longitudinal phenotyping to translate genetic discovery into better care for children with autism worldwide.

Subject of Research: Trio-based whole-exome sequencing to identify rare genetic variants in Rwandan children with autism spectrum disorder

Article Title: Clinical Utility of Trio Exome Sequencing in Rwandan Children With Autism Spectrum Disorder

Article References: Hakizimana, O., Hitayezu, J., Uyisenga, J. P., Dukuze, N., Akimana, M. V., Mizero, L., Bampire, C., Mudenge, C., Butoto, X. K., Helou, L., Charloteaux, B., Caberg, J.-H., Dideberg, V., Palmeira, L., Alagbonsi, A. I., Bours, V., & Uwineza, A. (2026). Clinical Utility of Trio Exome Sequencing in Rwandan Children With Autism Spectrum Disorder. Molecular Genetics & Genomic Medicine, 14(9), Article e70294. https://doi.org/10.1002/mgg3.70294

Image Credits: AI Generated

DOI: 10.1002/mgg3.70294

Keywords: autism spectrum disorder, whole-exome sequencing, Rwanda, trio sequencing, genetic variants, copy-number variants, de novo mutations, GABRB3, SYNGAP1, SHANK3, genomic medicine, African genomics

The Philadelphia Neurodevelopmental Cohort is a population-based sample from the greater Philadelphia area, originally comprising more than 9,400 young people. For the new analyses, the team restricted the sample to 4,694 white non-Hispanic participants with both genetic and cognitive data, a step taken to control for population stratification, a known source of false positives in genetic studies. The participants had a mean age of 13.8 years and were almost evenly split by sex. Each had completed the Penn Computerized Neurocognitive Battery, a 14-test instrument designed to probe five domains of cognition: executive function, encompassing abstraction and mental flexibility, attention, and working memory; episodic memory, covering verbal, facial, and spatial recall; complex cognition, including verbal reasoning, nonverbal reasoning, and spatial processing; social cognition, measured through tasks of emotion identification, emotion differentiation, and age differentiation; and speed, assessed through motor and sensorimotor reaction-time tasks. Both accuracy and response time were recorded for every test.

To make the data tractable for genetic analysis, the researchers compressed the battery into composite domain scores using confirmatory factor analysis, and derived overarching measures of general accuracy, commonly denoted “g,” and general speed, denoted “gs,” through principal component analysis. Missing cognitive data were handled with multiple imputation by chained equations, a standard statistical technique that fills gaps using age, sex, and available test results. Genotyping was performed on Illumina arrays and imputed against the 1000 Genomes reference panel at the Broad Institute, and the team estimated empirical genetic relatedness among all participants from roughly 50,000 common single-nucleotide polymorphisms. Including related individuals, the authors note, substantially boosts power to detect heritability and, critically, to detect Gene-by-Sex interactions.

The statistical machinery at the heart of the study is an extension of the classical polygenic model implemented in the SOLAR software package. Under a univariate model, the phenotypic variance of a trait is decomposed into additive genetic variance and environmental variance, and narrow-sense heritability is the ratio of the former to the total. The researchers then extended this model in two directions. The first is a quantitative Gene-by-Sex interaction: here, the additive genetic variance is allowed to differ between females and males, and the likelihood of that model is compared with one in which the genetic variances are constrained to be equal. A significant result indicates that genes matter more in one sex than the other, even if the same genes are involved. The second is a qualitative interaction, tested through the genetic correlation between the sexes. If this correlation falls below 1, it implies that different genetic variants influence the trait in females and males; if it equals 1, the genes are the same. All models adjusted for age, age squared, sex, and their interactions, and false discovery rate was controlled at 5 percent.

The behavioral findings, first, are strikingly consistent with decades of prior work. Differences between the sexes were small or negligible on nearly every measure, with standardized mean differences ranging from 0.061 to 0.182, all below the conventional threshold of 0.2 standard deviations that defines a “small” effect. Females showed higher accuracy on memory and social cognition tasks, and faster performance on complex and social cognition speed measures. Males showed higher accuracy on executive and complex cognition, and faster executive and sensorimotor speed. On the omnibus measures, there was no significant difference in general cognitive accuracy, but females were modestly faster on general speed. Notably, the largest single difference, in sensorimotor speed favoring males, remained small at 0.182 standard deviations, underscoring how little overlap-free separation there truly is between the sexes on these dimensions.

The genetic results, however, tell a more nuanced story. All cognitive accuracy measures were moderately heritable, with estimates ranging from 0.417 to 0.718, and all speed measures were heritable in the range of 0.363 to 0.471. But when the polygenic models were refit separately for each sex, the magnitude of genetic influence diverged in domain-specific ways. On executive function accuracy, males showed roughly twice the genetic variance of females, with estimates of 0.598 versus 0.301, and correspondingly lower environmental variance. On complex cognition accuracy the pattern repeated: genetic variance of 0.610 in males against 0.291 in females. Conversely, on complex cognition speed and social cognition speed, females carried the larger genetic burden, with genetic variance of 0.575 and 0.589 respectively, compared with 0.135 and 0.129 in males, whose performance on those speed measures was instead dominated by environmental variance exceeding 0.64. In short, the genes that matter appear to matter differently in different domains, and differently in each sex.

Yet the qualitative picture is one of remarkable unity. Genetic correlations between females and males were not significantly different from 1 on any of the cognitive measures. This suggests complete pleiotropy, the situation in which the same variants influence the trait in both sexes. The finding aligns with a broader literature showing that for most complex human traits, heritability does not differ by sex and genetic correlations between sexes hover at or near unity. It also echoes a recent sex-stratified genome-wide association study of brain anatomy that found largely consistent genetic influences across sexes. In the context of the new results, this means that sex differences in cognition are not a story of different genes at work, but of the same genetic architecture being amplified or muted by sex-specific biological and social contexts.

The authors offer several candidate mechanisms for why genetic variance might be larger in one sex than the other. One prominent hypothesis involves gene-environment correlation: individuals may select or evoke experiences that align with their genetic predispositions, and the degree to which this happens may itself be sex-dependent. Girls, for instance, may be more likely to elicit or seek out prosocial experiences that reinforce genetic tendencies in social cognition, amplifying genetic differences on that domain. Boys may do the same for activities that exercise spatial and executive skills. Hormonal factors are also on the table. Puberty brings substantial testosterone exposure, and prior work has linked pubertal testosterone to mental rotation performance in males, suggesting that sex hormones may magnify the expression of existing genetic influences during adolescence. Epigenetic mechanisms, in which sex-differential methylation or chromatin modification alters gene expression, provide a further layer of plausibility.

The study connects its findings to psychiatric epidemiology in a way that gives the results added clinical weight. Social cognitive deficits are documented features of schizophrenia, depression, and bipolar disorder, and the prevalence of these conditions differs markedly between females and males. If the genetic architecture of social cognition is more strongly expressed in females, as the new data suggest, then models linking social cognition to psychiatric risk may need to be parameterized by sex. Similarly, understanding why genetic influences on executive and complex cognition are more strongly expressed in males could illuminate why neurodevelopmental conditions with executive dysfunction show male-skewed prevalence. The authors caution, however, that their findings are correlational and population-level, and should not be read as deterministic predictions about any individual.

Limitations temper the conclusions. The sample was restricted to individuals of European ancestry, and the authors explicitly call for replication in African, Asian, and Hispanic populations, noting that their cohort included too few participants of other ancestries for stratified analysis. The cohort spans ages 8 to 21 but the models did not account for puberty staging, which given the hormonal hypotheses above is a meaningful gap. And while the sample is large by the standards of behavior genetics, the statistical power to detect genetic correlations below 1 is famously demanding: the authors calculate that even with 80 percent power, detecting a genetic correlation of 0.7 given a heritability of 70 percent requires a sample of this approximate size, while detecting a correlation of 0.9 would require more than 30,000 twins in a classical twin design. Smaller-than-perfect genetic correlations may therefore exist but evade detection.

What the study makes unambiguous is that the social-versus-biological framing that has long dominated discussion of cognitive sex differences is inadequate. The effect sizes for phenotypic differences are small, and no finding here should be mistaken for evidence that one sex is cognitively superior to the other. But the genetic variance findings, the authors argue, demonstrate that genetic factors are not a negligible or secondary consideration. Instead, genes, environment, hormones, and social context appear to operate jointly, with genetic influences being channeled, buffered, or amplified by sex-specific developmental pathways. Untangling that interplay, the researchers conclude, is the necessary next step for a field whose subjects, cognition and sex, sit at the intersection of psychology, neuroscience, genetics, and social science.

In the realm of neuroscience, progenitor cells have a pivotal role; they are the fundamental building blocks from which nerve cells develop. The study, led by Michael Heide and involving significant contributions from Nesil Eşiyok, focused on two human-specific genes: NBPF14 and NOTCH2NLB. These genes are instrumental in the regulation of progenitor cell proliferation and differentiation, which, in turn, affects the overall complexity and size of the brain. This finely tuned balance between cell multiplication and transformation is crucial during developmental phases, particularly in prenatal stages where the foundation for future cognitive capabilities is laid.

Understanding the cellular mechanisms underlying brain development has implications beyond just evolutionary biology; it opens doors to uncovering the etiology of various neurodevelopmental disorders that may arise due to disruptions in these processes. Disorders such as autism and schizophrenia have been associated with altered progenitor cell dynamics, hence investigating how these specific genes operate could provide valuable insights into their pathogenesis. Eşiyok articulates the significance of these findings, emphasizing that they offer a fresh perspective on both the evolution of the human brain and the potential for addressing developmental malformations that might stem from genetic anomalies.

Additionally, the hybrid methodological approach adopted in this research enhances the validity and richness of the findings. By integrating traditional animal model experiments with innovative alternatives like chimpanzee brain organoids, researchers were able to corroborate results across different systems. This synergistic strategy not only affirms the reliability of the conclusions drawn but also highlights an evolving paradigm in scientific research where alternative methods can complement and potentially reduce reliance on conventional animal testing.

Furthermore, the implications of this study resonate on a broader scale, particularly in the ongoing dialogue about ethical research practices and animal welfare. By advancing alternative methodologies, the scientific community is tasked with refining tools that could allow for comprehensive studies without the necessity of extensive animal trials. This not only aligns with contemporary ethical standards but also propels the field toward innovative solutions that could minimize the impact on live subjects while still obtaining crucial data.

The research positions itself within the larger context of neurobiology and evolutionary studies, linking genetic factors to observable traits in brain structure and function. As the researchers elucidate the specific roles of NBPF14 and NOTCH2NLB, they underscore the significance of these genes in the context of human evolutionary history. The research findings suggest that the unique attributes of the human brain, including its size and complexity, can be traced back to mutations and adaptations of these genes that have occurred over millennia.

Moreover, the article titled "A dyad of human-specific NBPF14 and NOTCH2NLB orchestrates cortical progenitor abundance crucial for human neocortex expansion," published in the reputable journal Science Advances, encapsulates the essence of this research, capturing the scientific community’s attention. The integration of various methods into one cohesive study presents a model for future research endeavors that seek to unravel the complexities of brain development while remaining cognizant of ethical concerns related to animal research.

The insights gleaned from this study not only push the boundaries of our understanding regarding human brain development but also stimulate ongoing discussions surrounding the biological underpinnings of human cognition. As we continue to explore the genetic influences on brain formation, the potential for groundbreaking therapeutic strategies rises—those that could specifically target the developmental pathways influenced by NBPF14 and NOTCH2NLB, ultimately offering hope for individuals affected by somatic mutations linked to neurodevelopmental disorders.

As the implications of the study echo throughout the scientific community, it provokes a renewed interest in interdisciplinary collaboration, uniting fields such as genetics, neuroscience, and evolutionary biology. The potential of this research to influence both academic discourse and practical therapeutic developments cannot be understated; it transcends mere academic curiosity and enters the domain of clinical applicability, establishing a robust foundation upon which future therapies could be built.

In conclusion, the findings from this research not only provide valuable contributions to our understanding of human brain evolution but also offer avenues for addressing existing challenges within the realm of developmental neurology. As contemporary research continues to unveil the interconnectedness of genetics and neurodevelopment, the work conducted at the German Primate Center stands as a testament to the power of scientific inquiry and its ability to illuminate the complexities of what makes us distinctly human.

Subject of Research: Cells
Article Title: A dyad of human-specific NBPF14 and NOTCH2NLB orchestrates cortical progenitor abundance crucial for human neocortex expansion
News Publication Date: October 2023
Web References: Science Advances
References: Not provided
Image Credits: Photo: Nesil Eşiyok

Keywords: Brain development, progenitor cells, neurobiology, genetics, human evolution, neurodevelopmental disorders, alternative research methods, ethical practices, animal testing reduction, neuroscience.