Lingo: an automated, web-based deep phenotyping platform for language ability

Background

Language and the ability to communicate effectively are key factors in mental health and well-being. Despite this critical importance, research on language is limited by the lack of a scalable phenotyping toolkit.

Methods

Here, we describe and showcase Lingo - a flexible online battery of language and nonverbal reasoning skills based on seven widely used tasks (COWAT, picture narration, vocal rhythm entrainment, rapid automatized naming, following directions, sentence repetition, and nonverbal reasoning). The current version of Lingo takes approximately 30 minutes to complete, is entirely open source, and allows for a wide variety of performance metrics to be extracted. We asked > 1,300 individuals from multiple samples to complete Lingo, then investigated the validity and utility of the resulting data.

Results

We conducted an exploratory factor analysis across 14 features derived from the seven assessments, identifying five factors. Four of the five factors showed acceptable test-retest reliability (Pearson's R > 0.7). Factor 2 showed the highest reliability (Pearson's R = 0.95) and loaded primarily on sentence repetition task performance. We validated Lingo with objective measures of language ability by comparing performance to gold-standard assessments: CELF-5 and the VABS-3. Factor 2 was significantly associated with the CELF-5 "core language ability" scale (Pearson's R = 0.77, p-value < 0.05) and the VABS-3 "communication" scale (Pearson's R = 0.74, p-value < 0.05). Factor 2 was positively associated with phenotypic and genetic measures of socieconomic status. Interestingly, we found the parents of children with language impairments had lower Factor 2 scores (p-value < 0.01). Finally, we found Lingo factor scores were significantly predictive of numerous psychiatric and neurodevelopmental conditions.

Conclusions

Together, these analyses support Lingo as a powerful platform for scalable deep phenotyping of language and other cognitive abilities. Additionally, exploratory analyses provide supporting evidence for the heritability of language ability and the complex relationship between mental health and language.

Genome-wide analyses of individual differences in quantitatively assessed reading- and language-related skills in up to 34,000 people

The use of spoken and written language is a fundamental human capacity. Individual differences in reading- and language-related skills are influenced by genetic variation, with twin-based heritability estimates of 30 to 80% depending on the trait. The genetic architecture is complex, heterogeneous, and multifactorial, but investigations of contributions of single-nucleotide polymorphisms (SNPs) were thus far underpowered. We present a multicohort genome-wide association study (GWAS) of five traits assessed individually using psychometric measures (word reading, nonword reading, spelling, phoneme awareness, and nonword repetition) in samples of 13,633 to 33,959 participants aged 5 to 26 y. We identified genome-wide significant association with word reading (rs11208009, P = 1.098 × 10-8) at a locus that has not been associated with intelligence or educational attainment. All five reading-/language-related traits showed robust SNP heritability, accounting for 13 to 26% of trait variability. Genomic structural equation modeling revealed a shared genetic factor explaining most of the variation in word/nonword reading, spelling, and phoneme awareness, which only partially overlapped with genetic variation contributing to nonword repetition, intelligence, and educational attainment. A multivariate GWAS of word/nonword reading, spelling, and phoneme awareness maximized power for follow-up investigation. Genetic correlation analysis with neuroimaging traits identified an association with the surface area of the banks of the left superior temporal sulcus, a brain region linked to the processing of spoken and written language. Heritability was enriched for genomic elements regulating gene expression in the fetal brain and in chromosomal regions that are depleted of Neanderthal variants. Together, these results provide avenues for deciphering the biological underpinnings of uniquely human traits.

Clinical autism subscales have common genetic liabilities that are heritable, pleiotropic, and generalizable to the general population

The complexity of autism's phenotypic spectra is well-known, yet most genetic research uses case-control status as the target trait. It is undetermined if autistic symptom domain severity underlying this heterogeneity is heritable and pleiotropic with other psychiatric and behavior traits in the same manner as autism case-control status. In N = 6064 autistic children in the SPARK cohort, we investigated the common genetic properties of twelve subscales from three clinical autism instruments measuring autistic traits: the Social Communication Questionnaire (SCQ), the Repetitive Behavior Scale-Revised (RBS-R), and the Developmental Coordination Disorder Questionnaire (DCDQ). Educational attainment polygenic scores (PGS) were significantly negatively correlated with eleven subscales, while ADHD and major depression PGS were positively correlated with ten and eight of the autism subscales, respectively. Loneliness and neuroticism PGS were also positively correlated with many subscales. Significant PGS by sex interactions were found-surprisingly, the autism case-control PGS was negatively correlated in females and had no strong correlation in males. SNP-heritability of the DCDQ subscales ranged from 0.04 to 0.08, RBS-R subscales ranged from 0.09 to 0.24, and SCQ subscales ranged from 0 to 0.12. GWAS in SPARK followed by estimation of polygenic scores (PGS) in the typically-developing ABCD cohort (N = 5285), revealed significant associations of RBS-R subscale PGS with autism-related behavioral traits, with several subscale PGS more strongly correlated than the autism case-control PGS. Overall, our analyses suggest that the clinical autism subscale traits show variability in SNP-heritability, PGS associations, and significant PGS by sex interactions, underscoring the heterogeneity in autistic traits at a genetic level. Furthermore, of the three instruments investigated, the RBS-R shows the greatest evidence of genetic signal in both (1) autistic samples (greater heritability) and (2) general population samples (strongest PGS associations).

Genetic and morphological estimates of androgen exposure predict social deficits in multiple neurodevelopmental disorder cohorts

BACKGROUND

Neurodevelopmental disorders (NDDs) such as autism spectrum disorder (ASD) display a strong male bias. Androgen exposure is profoundly increased in typical male development, but it also varies within the sexes, and previous work has sought to connect morphological proxies of androgen exposure, including digit ratio and facial morphology, to neurodevelopmental outcomes. The results of these studies have been mixed, and the relationships between androgen exposure and behavior remain unclear.

METHODS

Here, we measured both digit ratio masculinity (DRM) and facial landmark masculinity (FLM) in the same neurodevelopmental cohort (N = 763) and compared these proxies of androgen exposure to clinical and parent-reported features as well as polygenic risk scores.

RESULTS

We found that FLM was significantly associated with NDD diagnosis (ASD, ADHD, ID; all [Formula: see text]), while DRM was not. When testing for association with parent-reported problems, we found that both FLM and DRM were positively associated with concerns about social behavior ([Formula: see text], [Formula: see text]; [Formula: see text], [Formula: see text], respectively). Furthermore, we found evidence via polygenic risk scores (PRS) that DRM indexes masculinity via testosterone levels ([Formula: see text], [Formula: see text]), while FLM indexes masculinity through a negative relationship with sex hormone binding globulin (SHBG) levels ([Formula: see text], [Formula: see text]). Finally, using the SPARK cohort (N = 9419) we replicated the observed relationship between polygenic estimates of testosterone, SHBG, and social functioning ([Formula: see text], [Formula: see text], and [Formula: see text], [Formula: see text] for testosterone and SHBG, respectively). Remarkably, when considered over the extremes of each variable, these quantitative sex effects on social functioning were comparable to the effect of binary sex itself (binary male: [Formula: see text]; testosterone: [Formula: see text] from 0.1%-ile to 99.9%-ile; SHBG: [Formula: see text] from 0.1%-ile to 99.9%-ile).

LIMITATIONS

In the devGenes and SPARK cohorts, our analyses rely on indirect, rather than direct measurement of androgens and related molecules.

CONCLUSIONS

These findings and their replication in the large SPARK cohort lend support to the hypothesis that increasing net androgen exposure diminishes capacity for social functioning in both males and females.

Estimating the Prevalence and Genetic Risk Mechanisms of ARFID in a Large Autism Cohort

This study is the first genetically-informed investigation of avoidant/restrictive food intake disorder (ARFID), an eating disorder that profoundly impacts quality of life for those affected. ARFID is highly comorbid with autism, and we provide the first estimate of its prevalence in a large and phenotypically diverse autism cohort (a subsample of the SPARK study, N = 5,157 probands). This estimate, 21% (at a balanced accuracy 80%), is at the upper end of previous estimates from studies based on clinical samples, suggesting under-diagnosis and potentially lack of awareness among caretakers and clinicians. Although some studies suggest a decrease of disordered eating symptoms by age 6, our estimates indicate that up to 17% (at a balanced accuracy 87%) of parents of autistic children are also at heightened risk for ARFID, suggesting a lifelong risk for disordered eating. We were also able to provide the first estimates of narrow-sense heritability (h²) for ARFID risk, at 0.45. Genome-wide association revealed a single hit near ZSWIM6, a gene previously implicated in neurodevelopmental conditions. While, the current sample was not well-powered for GWAS, effect size and heritability estimates allowed us to project the sample sizes necessary to more robustly discover ARFID-linked loci via common variants. Further genetic analysis using polygenic risk scores (PRS) affirmed genetic links to autism as well as neuroticism and metabolic syndrome.

Forecasting risk gene discovery in autism with machine learning and genome-scale data

Genetics has been one of the most powerful windows into the biology of autism spectrum disorder (ASD). It is estimated that a thousand or more genes may confer risk for ASD when functionally perturbed, however, only around 100 genes currently have sufficient evidence to be considered true "autism risk genes". Massive genetic studies are currently underway producing data to implicate additional genes. This approach - although necessary - is costly and slow-moving, making identification of putative ASD risk genes with existing data vital. Here, we approach autism risk gene discovery as a machine learning problem, rather than a genetic association problem, by using genome-scale data as predictors to identify new genes with similar properties to established autism risk genes. This ensemble method, forecASD, integrates brain gene expression, heterogeneous network data, and previous gene-level predictors of autism association into an ensemble classifier that yields a single score indexing evidence of each gene's involvement in the etiology of autism. We demonstrate that forecASD has substantially better performance than previous predictors of autism association in three independent trio-based sequencing studies. Studying forecASD prioritized genes, we show that forecASD is a robust indicator of a gene's involvement in ASD etiology, with diverse applications to gene discovery, differential expression analysis, eQTL prioritization, and pathway enrichment analysis.

Whole-genome sequencing in a family with twin boys with autism and intellectual disability suggests multimodal polygenic risk

Over the past decade, a focus on de novo mutations has rapidly accelerated gene discovery in autism spectrum disorder (ASD), intellectual disability (ID), and other neurodevelopmental disorders (NDDs). However, recent studies suggest that only a minority of cases are attributable to de novo mutations, and instead these disorders often result from an accumulation of various forms of genetic risk. Consequently, we adopted an inclusive approach to investigate the genetic risk contributing to a case of male monozygotic twins with ASD and ID. At the time of the study, the probands were 7 yr old and largely nonverbal. Medical records indicated a history of motor delays, sleep difficulties, and significant cognitive deficits. Through whole-genome sequencing of the probands and their parents, we uncovered elevated common polygenic risk, a coding de novo point mutation in CENPE, an ultra-rare homozygous regulatory variant in ANK3, inherited rare variants in NRXN3, and a maternally inherited X-linked deletion situated in a noncoding regulatory region between ZNF81 and ZNF182 Although each of these genes has been directly or indirectly associated with NDDs, evidence suggests that no single variant adequately explains the probands' phenotype. Instead, we propose that the probands' condition is due to the confluence of multiple rare variants in the context of a high-risk genetic background. This case emphasizes the multifactorial nature of genetic risk underlying most instances of NDDs and aligns with the "female protective model" of ASD.

cerebroViz: an R package for anatomical visualization of spatiotemporal brain data

Summary

Spatiotemporal transcriptomic profiling has provided valuable insight into the patterning of gene expression throughout the human brain from early fetal development to adulthood. When combined with prior knowledge of a disease's age at onset and region-specificity, these expression profiles have provided the necessary context to both strengthen putative gene-disease associations and infer new associations. While a wealth of spatiotemporal expression data exists, there are currently no tools available to visualize this data within the anatomical context of the brain, thus limiting the intuitive interpretation of many such findings. We present cerebroViz, an R package to map spatiotemporal brain data to vector graphic diagrams of the human brain. Our tool allows rapid generation of publication-quality figures that highlight spatiotemporal trends in the input data, while striking a balance between usability and customization. cerebroViz is generalizable to any data quantifiable at a brain-regional resolution and currently supports visualization of up to thirty regions of the brain found in databases such as BrainSpan, GTEx and Roadmap Epigenomics.

Availability and Implementation

cerebroViz is freely available through GitHub ( https://github.com/ethanbahl/cerebroViz ). The tutorial is available at http://ethanbahl.github.io/cerebroViz/.

Contacts

ethan-bahl@uiowa.edu or jacob-michaelson@uiowa.edu.

Supplementary information

Supplementary data are available at Bioinformatics online.