Towards Disentangling the Polygenic Contribution of Dyslexia to School Performance

Towards Disentangling the Polygenic Contribution of Dyslexia to School Performance | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Towards Disentangling the Polygenic Contribution of Dyslexia to School Performance Judit Cabana-Domínguez, Marta Graell, Rosa Bosch, María Soler Artigas, and 15 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7206327/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Dyslexia is a neurodevelopmental disorder typically diagnosed in school-aged children and associated with poor school performance and lower levels of educational attainment (EA). Here, we aimed to test the shared genetic architecture between dyslexia and EA, to dissect the polygenic contribution of dyslexia by its relationship with EA and to assess how these genetic partitions influence school performance, early manifestation of psychopathology and related traits. We first confirmed a negative genetic correlation between dyslexia and EA (rg=-0.186, SE = 0.019, P = 1.75E-22). Then, polygenic scores for EA and dyslexia were tested in a cohort of 4,274 school-aged children, revealing opposite direction of the effect in school performance. Next, we dissected the genetic liability for dyslexia into components shared with, and independent of, EA. The results revealed similar patterns of association for performance in primary and foreign languages, but distinct patterns when comparing these language-related subjects with mathematics. The dyslexia-specific genetic component independent of EA was associated with poorer academic outcomes in language-related subjects and increased rates of psychopathology, supporting the existence of dyslexia-specific genetic effects beyond general cognitive or educational pathways. In contrast, the genetic load of dyslexia that overlaps with EA contribute to school performance in both language-related subjects and mathematics and displayed opposite patterns of association dependent on whether concordant and discordant genomic partitions were considered. The discordant partition was associated with poorer school performance and higher rates of behavioral and emotional problems, being these associations partially mediated by the dyslexia diagnosis (accounting for a reduction in effect size ranging from 10.44 to 12.91%). Conversely, the concordant partition was only associated with better performance in mathematics. Overall, these findings highlight the polygenic contribution of dyslexia to both academic and psychopathological outcomes, support distinct genetic influences on language skills and mathematics, and uncerscore the usage of the genetic load for EA to deepen insight into the complex genetic relationship between dyslexia and school performance. Biological sciences/Genetics/Genomics Health sciences/Diseases/Psychiatric disorders learning disorder educational attainment polygenic risk scores shared genetics Figures Figure 1 Figure 2 Figure 3 Figure 4 INTRODUCTION Dyslexia is a neurodevelopmental disorder commonly diagnosed in school-aged children, characterized by a persistent impairment in specific learning skills such as reading accuracy, spelling, grammar, or reading fluency. It is found worldwide and across all languages, including different alphabetic orthographies and logographic languages 1 , with prevalence estimates ranging from 5–20% 2 . Dyslexia typically manifests in educational settings, but its impacts often extends beyond the classroom, affecting self-esteem, social interactions, and long-term occupational prospects 3 . Additionally, dyslexia frequently co-occurs with other psychiatric and neurodevelopmental disorders, including attention-deficit/hyperactivity disorder (ADHD) and language disorders 4 . Dyslexia is a complex disorder with an estimated heritability of up to 70% from twin and family studies 5 . The largest genome-wide association study (GWAS) on dyslexia to date (51,800 cases and 1,087,070 controls) identified 42 independent hits and significant genetic correlations between dyslexia and 63 phenotypes including several mental disorders and behavioral traits (e.g., ADHD and risk-taking behavior) as well as measures of academic performance (e.g., educational attainment (EA) and age completed education) 6 . Dyslexia is related with low EA 7 – 9 , which represents the highest level of education an individual has completed. EA is commonly used as proxy to quantify educational performance and is also linked to different developmental disorders, as well as social, economic and health outcomes 10 – 12 . Similar to dyslexia, EA is a highly polygenic trait, with 1,271 independent hits identified in one of the largest GWAS published to date in about 1.1 million individuals 10 . Dyslexia is also negatively associated with school performance, which is considered an early-life intermediate phenotype that may anticipate EA 13 , 14 . Children with dyslexia have an increased risk of obtaining lower grades compared to other peers and, consequently, a greater risk of falling behind on, or dropping out from, school 15 , 16 . These academic challenges can result in a negative feedback loop, worsening the self-confidence and motivation of these students, further aggravating their academic difficulties 17 . Furthermore, the stigma around learning disabilities can lead to social isolation, contributing to a negative school experience 18 . Dyslexia and EA are complex and polygenic traits, but their genetic relationship with school performance remains underexplored. We aimed to clarify whether the association between the genetic liability of dyslexia with school performance is partially explained by the genetic liability for EA. Using a deeply phenotyped cohort of 4,274 school-aged children, we sought to: (i) assess the shared genetic architecture between dyslexia and EA; (ii) explore the association of genetic liability of dyslexia and EA with school performance; (iii) determine whether the genetic contribution of dyslexia to school performance is influenced by the genetic liability of EA; and (iv) evaluate the association of dyslexia with childhood psychopathology and other psychiatric disorders and related traits, as well as assess the role of EA in these associations (Supplementary Figure S1 ). MATERIALS AND METHODS Study sample, school performance and clinical assessment The study sample from the INSchool cohort included 4,274 school-aged participants (mean age = 10.03 years, SD = 2.95, ranging from 5 to 17 years; 43.7% females), representative of the general school-aged population, recruited in 45 primary and secondary schools across Catalonia, Spain 19 . The inclusion criteria for participants were European ancestry and written informed consent from parents or caregivers. School performance data was provided by schools and assessed with school grades in three subjects: mathematics, primary language and foreign language. Grades were described on a four-point scale, being A: excellent performance, B: good performance, C: adequate performance and D: underperformance. Dyslexia was diagnosed in 9.66% of the 4,274 participants (n = 413). Assessment was conducted using the Battery for the Evaluation of Reading Processes, Revised (PROLEC-R) 20 and the Battery for the Evaluation of Reading Processes in Junior and Senior High-School Students, Revised (PROLEC-SE-R) 21 . Participants with intellectual disability were excluded based on results from the Wechsler Intelligence Scale for Children (WISC) 22 , as previously described 23 . Childhood psychopathology was assessed using the Child Behavior Checklist for ages 6–18 (CBCL-6/18) from Achenbach System of Empirically Based Assessment (ASEBA) 24 completed by parents or surrogates in most of the school-aged sample (96%; n = 4,084). Genotyping, imputation and quality control Genomic DNA isolated from saliva samples or buccal swabs was genotyped in three different waves with the Illumina Infinium PsychChip_v1.0 array for wave 1 (n = 793) or the Infinium Global Screening Array-24 version_2 (GSA_v2) for waves 2 (n = 2732) and 3 (n = 749) (Illumina, CA, San Diego, USA). Pre-imputation quality control was done with PLINK_2.0 software as described in Cabana-Domínguez et al. 2024 25 , prepared for imputation using McCarthy tools, and imputed with the Michigan Imputation Server separately 26 . After imputation, we only kept SNPs with an imputation INFO score > 0.8, minor allele frequency (MAF) > 0.01 and SNP call rate > 0.95. Then, samples from the three GWAS waves were merged and only SNPs present in all of them, with MAF > 0.01, SNP call rate > 0.95 and differential case/control missingness rate < 0.2 in the overall sample were considered for subsequent analyses, ending up with 3,725,956 SNPs. GWAS summary statistics Summary statistics from GWAS meta-analysis on EA (N = 1,131,881 individuals) 10 and dyslexia (N = 51,800 cases and 1,087,070 controls, N effective =197,776 individuals) 6 were obtained from 23andMe, Inc. The effective sample size for dyslexia was calculated as N effective = 4/(1/N cases + 1/N controls ). SNPs with minor allele frequency (MAF) < 0.01 or SNPs located on the X, Y and mitochondrial chromosomes were filtered out. Genome-wide polygenic scores (PGS) PGS were constructed in our in-house sample of 4,274 school-aged participants using summary statistics of EA and dyslexia with the PRS-CS 27 and PLINK_2.0 softwares. Scores were calculated using the common set of SNPs between EA and dyslexia (N commonSNP =6,429,800). Once computed, all PGS were standardized to a mean of 0 and a standard deviation of 1. Additionally, PGS for dyslexia (PGS DYS ) were constructed with four subsets of SNPs based of their contribution to EA, as previously described 25 , 28 . Prior to subsetting, GWAS summary statistics were harmonized to ensure alignment of effect alleles across both studies. Then, genetic liability of dyslexia was divided in hierarchical subsets as follows (Supplementary Figure S1 ): (i) variants not associated with EA (PGS DYS_noEA ; P EA >0.05; N SNPs = 5,004,441), and (ii) variants associated with EA (PGS DYS_EA ; P EA ≤0.05; N SNPs = 1,425,359). The second group was divided in two additional subsets based on the direction of effects in dyslexia and EA: (iii) variants showing consistent direction of effect in both dyslexia and EA (PGS DYSconcordant ; Beta DYS >0 and Beta EA >0 or Beta DYS <0 and Beta EA 0 and Beta EA <0 or Beta DYS 0; N SNPs = 789,169). Statistical analyses SNP-based heritability and genetic overlap Summary statistics from GWAS meta-analysis on EA 10 and dyslexia 6 were used to estimate the SNP-based heritability (h 2 SNP ) and the genetic correlation between EA and dyslexia using LDSC_v1.0 29 . All analyses were restricted to Hapmap3 SNPs and considering a population prevalence for dyslexia of 10% 6 . In addition, MiXeR ( https://github.com/precimed/mixer ) 30 was used to calculate the number of trait-influencing SNPs with the univariate model, and the genetic overlap between traits with the bivariate model. The Dice coefficient explains the proportion of shared SNPs between traits, and the Akaike Information Criterion (AIC) was used to determine the model fit ( https://github.com/precimed/mixer#aic-bic-interpretation ). Partitioned heritability SNP based partitioned heritability (h 2 SNP ) was estimated for the three independent dyslexia partitioned subsets (DYS noEA , DYS concordant and DYS discordant ) and calculated using LDSC_v1.0 31 . The heritability was computed for each partion using the GWAS summary statistics of dyslexia 6 and data from the 1000G as reference panel, and the analysis was restricted to Hapmap3 SNPs. Enrichment of genome-wide significant hits (P < 5e-08) of dyslexia in the genome partitions described previously were calculated using a Chi-square test and Manhattan plots were obtained using the qqman R package. Association between PGS and school performance PGS were associated with school performance using an ordinal mixed-effect model considering A as the highest and D as the lowest category using the ordinal R package 32 . The percentage of variance attributable to each PGS was calculated as the increase in Naggelkerke’s pseudo-r2 between models with and without the PGS. P-values were corrected for multiple comparisons via Bonferroni correction. To compare the effect of different PGS, the target sample was divided into five quintiles of increasing PGS and odds ratio were compared using the lowest quintile as reference and represented in quantile plots. The predicted probabilities of being in each category (A, B, C or D) were calculated for each subject in order to visualize how the PGS affects school performance and represented in probability plots using the effects R library 33 . Sex-stratified analyses were conducted to examine whether the association between PGS and school performance differed between males and females. The same regression model described for the whole sample was followed. The regression model was modified to include the interaction term PGS*sex, allowing the effect of the PGS on school performance to vary by sex and capture any differential impact. The association between PGS and behavioral and emotional problems was tested using linear-mixed effects models. The CBCL-6/18 parent rating scales were used as continuous variables and square-root transformed because of skewness. P-values were corrected for multiple comparisons via the Benjamini-Hochberg False Discovery Rate (FDR) method (P adj <0.05). All analyses were adjusted for age, sex, socio-economic status (SES) and 20 genetic PC as fixed effects, as well as school as random effect to account for the multilevel nature of the data. The regression analyses for the quantile plots could not include school as a covariate, given the little variability for this variable in the smaller subsets of the data used for this analysis. SES was calculated using the Hollingshead Four-Factor Index based on parent’s education and occupation 34 . Mediation analysis We investigated whether the effect of the PGS on school performance was mediated by the diagnosis of dyslexia in the INSchool cohort. The effect of PGS on the mediator was estimated using a logarithmic mixed-effect logistic regression model, and the effect of the mediator on school performance was estimated using an ordinal mixed-effect model. When these two models had significant associations (P < 0.05), the effect of the mediator and the direct effect of the PGS on school performance was estimated using ordinal mixed-effect model. All analyses were adjusted for age, sex, SES and 20 genetic PC as fixed effects, as well as school as random effect. Partial mediation was determined when the direct effect remained significant but attenuated. Genetic covariance We used annotation-stratified genetic covariance analysis to examine how three partitioned subsets of dyslexia-associated SNPs (DYS noEA , DYS concordant , and DYS discordant ) contribute to the genetic overlap between dyslexia and other traits. This analysis was conducted using the GeNetic cOVariance Analyzer (GNOVA) ( https://github.com/xtonyjiang/GNOVA ) 35 , which estimates genetic covariance between two phenotypes while accounting for linkage disequilibrium (LD) and sample overlap. We included GWAS summary statistics for nine cognitive and language-related traits, seven psychiatric disorders, three substance use-related traits, and five traits related to well-being and socio-economic status (see Supplementary Table 1 for details). Genetic covariance was calculated using partial correlations, restricted to SNPs within each of the three dyslexia subsets, as well as across the full set of SNPs for dyslexia and educational attainment (EA). Following the developer’s recommendation, genomic covariance instead of correlation estimates were considered in all the analyses and P-values were adjusted for multiple comparisons using the FDR (P adj < 0.05). RESULTS Genetic architecture of dyslexia and EA We estimated a SNP-based heritability (h 2 SNP ) of 0.179 for dyslexia (SE = 0.0071) and 0.102 for EA (SE = 0.0023) and found a negative genetic correlation between them (rg=-0.186, SE = 0.019, P = 1.75E-22) (Supplementary Table 2). MiXeR analysis revealed that both traits are highly polygenic, with 10,315 common variants for dyslexia and 14,403 for EA. Most of the variants that influence dyslexia are shared with EA (99.5%; n = 10,214, SE = 465, Dice coefficient = 0.826), with a slightly higher proportion of variants with discordant direction of the effect between them (58.0%) (Fig. 1 and Supplementary Table 3). The genetic correlation was consistent with the stratified cross-phenotype Q-Q plots, where P-values for one trait are plotted conditioned on different association strengths for the other trait (Fig. 1 ). Contribution of the genetic liability for dyslexia and EA on school performance Genome-wide polygenic scores for EA (PGS EA ) were associated with higher odds of having a better performance in all three subjects studied in the 4,274 school-aged participants from the INSchool cohort (OR > 1.38, P < 4.41E-24). In contrast, PGS for dyslexia (PGS DYS ) were associated with worse performance across all three subjects, being the effect greater in language-related subjects (first language (OR = 0.79, 95% CI = 0.74–0.84) and foreign language (OR = 0.80, 95% CI = 0.75–0.85)) than in mathematics (OR = 0.89, 95% CI = 0.84–0.94). The proportion of variance in school performance explained by the genetic liability of EA ranges from 2.69–3.88% across subjects, while the genetic liability of dyslexia accounts for 0.46–1.65% (Table 1 and Fig. 2 ). Table 1 Association between school performance and genome-wide polygenic score for dyslexia (PGSDYS) based on its relationship with educational attainment (EA). Subject PGS OR (95% CI) Statistic P-value # R 2 * First Language EA 1.43 (1.34–1.53) 11.006 3.59E-28 3.23% DYS 0.79 (0.74–0.84) -7.873 3.45E-15 1.65% DYS no EA 0.84 (0.79–0.89) -5.887 3.93E-09 0.92% DYS - EA 0.81 (0.76–0.86) -7.141 9.29E-13 1.36% DYS concordant 1.06 (1-1.13) 2.020 0.044 0.11% DYS discordant 0.73 (0.69–0.78) -9.984 1.79E-23 2.65% Foreign Language EA 1.38 (1.30–1.47) 10.122 4.41E-24 2.69% DYS 0.80 (0.75–0.85) -7.637 2.23E-14 1.53% DYS no EA 0.85 (0.8–0.9) -5.621 1.90E-08 0.83% DYS - EA 0.82 (0.77–0.86) -6.952 3.61E-12 1.27% DYS concordant 1.04 (0.98–1.1) 1.356 0.175 0.05% DYS discordant 0.76 (0.72–0.81) -9.085 1.04E-19 2.16% Mathematics EA 1.47 (1.38–1.57) 12.210 2.75E-34 3.88% DYS 0.89 (0.84–0.94) -4.191 2.78E-05 0.46% DYS no EA 0.94 (0.89-1) -2.026 0.043 0.11% DYS - EA 0.87 (0.82–0.92) -4.983 6.27E-07 0.65% DYS concordant 1.10 (1.04–1.16) 3.185 1.45E-03 0.26% DYS discordant 0.78 (0.74–0.83) -8.227 1.92E-16 1.76% # P-values surpassing Bonferroni correction (P-value < 0.05/18 = 2.7E-03) are shown in bold. * Explained variance attributable to PGS calculated as the increase in Nagelkerke’s pseudo-R2 between an ordinal mixed-effect model with and without the PGS variable. PGS comparisons across ranked quintiles showed that the odds for better performance in children at the highest quintile for PGS EA was, on average, more than twice that in children in the first quintile for the three subjects. In contrast, we found the opposite effect of PGS DYS , where children in the highest quantile for PGS DYS have worse performance than children in the first quantile (Fig. 2 and Supplementary Table 4). In mathematics, we found that only the subset of individuals with the highest PGS DYS (5th quantile) show significant differences from those in the 1st quantile. These results are consistent with the increased probability of reaching better performance (grades A or B) with higher PGS EA or lower PGS DYS , and worse performance (grades C or D) with higher PGS DYS or lower PGS EA (Supplementary Figure S2 ). Polygenic dissection of the contribution of dyslexia to school performance by their relationship with EA The genetic load of dyslexia was divided in variants not associated with EA (DYS noEA ) and variants associated with EA (DYS EA ), which were then divided into variants with a consistent effect (DYS concordant ) and variants with opposing effect (DYS discordant ) (Supplementary Figure S1 ). Partitioned heritability estimates were calculated for the non-overlapping subsets of variants (DYS noEA , DYS discordant and DYS concordant ). We found that most of the h 2 SNP of dyslexia was distributed between DYS noEA and DYS discordant , being the last one especially relevant considering that a small proportion of SNPs (16.8%) explains a considerable portion (29.1%) of the h 2 SNP of dyslexia (Supplementary Table 5A). In addition, genome-wide significant variants for dyslexia identified by Doust et al. 6 were enriched across all analyzed subsets (Supplementary Figure S3 and Table 5B). PGS for dyslexia were calculated for the four subsets of genetic variants and tested for association with school performance in our in-house cohort of 4,274 school-aged children and adolescents. The results revealed similar patterns of association for performance in primary and foreign languages, but distinct patterns when comparing these language-related subjects with mathematics. PGS for dyslexia excluding EA-associated variants, PGS DYS_noEA , were associated with worse performance in first language (OR = 0.84, 95% CI = 0.79–0.89) and foreign language (OR = 0.85, 95% CI = 0.8–0.9), but showed no significant association with performance in mathematics. In contrast, PGS for dyslexia including EA-associated variants, PGS DYS_EA , were associated with worse performance across all three subjects (OR < 0.94, adj-P ≤ 0.046), although the proportion of variance explained was greater for language-related subjects (first language: 1,36% and foreign language: 1.27%) than for mathematics (0.65%) (Table 1 and Fig. 2 ). When dissecting variants shared with EA, we found that PGS DYSdiscordant was associated with worse performance in all subjects (OR < 0.78, adj-P ≤ 8.8E-15) and explained the largest proportion of variance on school performance of all genetic partitions, ranging from 1.76–2.65% (Table 1 and Fig. 2 ). Conversely, PGS DYSconcordant was associated with better performance in mathematics (OR = 1.10, 95% CI = 1.04–1.06), but dispayed no significant association in language-related subjects (Table 1 and Fig. 2 ). PGS comparisons across ranked quintiles showed the expected trend of lower odds for better school performance in individuals with higher PGS DYS across all partitions but PGS DYSconcordant (Fig. 2 and Supplementary Table 6). This result is consistent with the probability plots for performance, where the PGS DYSconcordant partition tends to the same effect as the PGS EA (i.e., probability of better performance (grades A or B) increase with higher PGS DYSconcordant ) while PGS DYSdiscordant was consistent with the results of PGS DYS (i.e., probability of worse performance (grades C or D) increased with higher PGS DYSdiscordant ) (Supplementary Figure S4). When stratifying by sex, no differences were found on the effect of the genetic liability of dyslexia dissected by its relationship with EA on school performance (Interaction P-value > 0.05) (Supplementary Table 7). To assess whether the dyslexia diagnosis explained the association of the different PGS with school performance, we first confirmed in our in-house sample of 4,274 school-aged participants, including 469 diagnosed cases, that the dyslexia diagnosis was associated with PGS DYSnoEA (P = 3.26E-04) and PGS DYSdiscordant (P = 4.25E-09), but not with PGS DYSconcordant , which were subsequently considered for the mediation analysis (Supplementary Table 8A). We also found association between the dyslexia diagnosis and school performance in the three subjects under study (P≤1.07E-39) (Supplementary Table 8B), and confirmed that the effect of both PGS DYSnoEA and PGS DYSdiscordant on school performance were attenuated after considering the effect of the dyslexia diagnosis. These results suggest that dyslexia partially mediates the effect of these PGS on school performance (accounting for a reduction in effect size ranging from 10.44–12.91%; Supplementary Tables 8C-D and Figure S5). Polygenic dissection of the contribution of dyslexia to childhood psychopathology by its relationship with EA The analysis of childhood psychopathology in our in-house INSchool cohort (n = 4084 school-aged individuals with completed clinical information) revealed that PGS EA was negatively associated with attentional problems and externalizing behavior. In contrast, the genetic liability of dyslexia and all its SNPs subsets, but the concordant genomic partition, were positively associated with somatic complains as well as social and attentional problems. In addition, the SNP subset with discordant direction of the effect between dyslexia and EA was also associated with externalizing behavior, indicating that this genomic partition is associated with the majority of behavioral and emotional problems assessed (Fig. 3 and Supplementary Table 9). Genetic covariance between dyslexia and related disorders and traits The analysis of genetic covariance showed opposite directions of effect for concordant and discordant genomic partitions across most of the assessed phenotypes. The patterns observed for the DYS discordant and DYS noEA partitions closely resembled the genetic correlations of dyslexia across the whole genome 6 . In contrast, the concordant genomic partition, DYS concordant , exhibited a pattern more aligned with the genetic correlations observed for EA. We found that all traits related with cognition and language ability- except for ambidextry- as well as anorexia nervosa, autism spectrum disorder (ASD), substance use related traits (i.e., lifetime cannabis use and alcohol use), subjective well-being, and household income, displayed negative covariance with the discordant but positive covariance with the concordant genetic load for dyslexia. In contrast, several psychiatric disorders (i.e., ADHD, major depression, and anxiety disorder), along with the addiction factor, sleeplessness/insomnia and loneliness/isolation, showed positive covariance with the discordant and negative covariance with the concordant genetic variation. Interestingly, significant genetic covariance with dyslexia was observed for anxiety disorder and lifetime cannabis use only when the genetic liablity was dissected by genetic variants associated with EA (i.e., DYS discordant and DYS concordant ). Finally, we also found that the genetic covariance of dyslexia with some traits was not significant for some genomic partitions (i.e., ambidexterity, schizophrenia and bipolar disorder), and for general risk tolerance was independent of EA and in the same direction than those described for dyslexia across the whole genome (Fig. 4 and Supplementary Table 10). DISCUSSION The results of this study confirmed negative genetic correlation between dyslexia and EA 7 – 9 and strong genetic overlap between them. Our findings also support the use of EA to dissect the genetic load of dyslexia, as previously described for schizophrenia 36 , ADHD and ASD 25 , to better understand the polygenic contribution of dyslexia to school performance and childhood psycopathology. We found that the genetic liability of EA and dyslexia were associated with school performance and showed opposite direction of the effect, as expected. PGS EA was strongly associated with better performance in primary language, foreign language and mathematics, while PGS DYS was associated with worse performance in all three subjects, as previously described in the literature 6 , 10 , 14 , 37 . Accordingly, the association between the PGS EA and PGS DYS with childhood psychopathology showcased this opposite pattern of association, with dyslexia being associated with more somatic complaints and social and attentional problems, some of them previously described 38 . When dissecting the genetic liability of dyslexia by its relationship with EA, we found that all genomic partitions were associated with school performance in some of the studied academic subjects. In line with the negative genetic correlation found between dyslexia and EA, the discordant genetic partition, which accounts for the highest enrichment in dyslexia genetic background 6 , showed the strongest negative effect in school performance, which was partially mediated by the dyslexia diagnosis. In addition, this genomic partition was associated with higher rates of children psychopathology, with special mention to externalizing behavior. These results confirm that a high proportion of dyslexia risk loci is shared with EA and has opposite direction of the effect, and provide further evidence on the negative effects of dyslexia on long-term academic and emotional outcomes 39 . Conversely, PGS constructed with concordant variants were associated with better performance in mathematics. This genomic partition showed notably low SNP heritability for dyslexia given the proportion of SNPs included, was not associated with dyslexia diagnosis nor did it contribute to language skills or children's psychopathology in our sample, which supports that this partition might be more closely related to EA rather than to dyslexia itself. This is further supported by the results in genetic covariance, where the concordant genomic partition showed greater similarity to the pattern observed in EA. Notably, most cognitive and psychiatric traits also exhibited opposite pattern of genetic covariance between the concordant and discordant genomic partitions of dyslexia, being the latter generaly associated with increased risk of mental disorders, reduced well-being, and lower cognitive performance, similarly to overall dyslexia variants described 6 . These findings are consistent with recent studies showing a connection between poor academic achievement and a higher risk of subsequent mental disorders, with the exception of eating disorders, where the association is of similar magnitude but in the opposite direction 40 . We also found that both anxiety and lifetime cannabis use were associated with the polygenic load of dyslexia only when its genetic liability was dissected by its relationship with EA. These findings suggest that these traits are highly influenced by educational level and are consistent with previous studies exploring the relationship between dyslexia and anxiety 41 . Overall, they reinforce the utility of EA-stratified analyses to uncover distinct genetic pathways within dyslexia and its comorbidities, and emphasize the importance of considering EA when studying the shared genetic underpinnings of cognitive and psychiatric outcomes. 6 We also found that the genetic liability of dyslexia not associated with EA has a negative impact on school performance, specially in first and foreing language. This suggests that the association between dyslexia and worse performance in language-related subjects is not only explained by shared genetic with EA, but also to specific dyslexia genetic liability independent of EA. Notably, this association was also partially mediated by dyslexia diagnosis and associated with higher rates of childhood psychopathology in our sample, following a pattern of association similar to the discordant genomic partition. These findings suggest that specific genetic components of dyslexia contribute to both language abilities and early psychopathologcal symptoms, beyond general cognitive or educational factors. This study aims to disentangle the polygenic contribution of dyslexia on school performance by incorporating a third trait, EA, facilitating a more comprehensive understanding of their polygenic effects and the early emergence of psychopathology in a well-characterized cohort. However, several considerations need to be taken into account. First, we observed that discordant polygenic signature between dyslexia and EA was associated with early-onset psychopathology and lower academic performance. This supports the existence of shared genetic factors that transcend diagnostic categories and suggests that early emerging symptoms may negatively impact school outcomes. However, we did not infer causality and cannot exclude reverse causation, where poor academic performance may itself increase the risk of developing psychopathology later in life 42 . This is specially relevant considering that there is evidence indicating that EA and school performance are influenced by demographic and indirect genetic effects 43 , such as assortative mating and dynastic effects, which may lead to inaccurate estimations of direct genetic effects 44 , 45 . While the primary goal of this study was to explore the polygenic contributions of dyslexia to school performance in relation to EA, our findings should be cautiously interpreted and further family-based studies accounting for these population mechanisms are required. Secondly, despite accounting for confounding factors, residual confounders may persist. Diverse dyslexia treatments across schools, ranging from cognitive and behavioral therapies to classroom support and individualized educational plans 46 , could influence school grades and potentially lead to an underestimation of polygenic effect of dyslexia on school performance. Lastly, analyses across the three academic subjects revelaed distinct patterns of association between language skills and mathematics. We found a stronger effect of the genetic load for dyslexia on language performance than on mathematics. Additionally, the genetic contribution on mathematics emerged only when considering the genomic patition shared with EA, whereas the dyslexia-specific genetic partition independent of EA contributed exclusively to school performance in language-related subjects. These findings support the existence of distinct genetic influences underlying language and mathematical abilities, highlighting the need for further studies to clarify these specific genetic contributions. In summary, we confirmed a negative genetic correlation between dyslexia and EA. By dissecting the genetic predisposition to dyslexia by its effect on EA, we found that the dyslexia genetic liability independent of EA was associated with poorer performance in language-related subjects and increased childhood psychopathology, supporting the existence of dyslexia-specific genetic effects beyond general cognitive or educational pathways. Moreover, the dyslexia genetic components associated with EA exhibited opposing effects: the concordant partition was associated with better performance only in mathematics, while the discordant partition was associated with worse school grades in all the studied subjects, partially mediated by dyslexia diagnosis, higher rates of children psychopathology, reduced well-being, and lower cognitive performance - mirroring the overall dyslexia signal. These findings reinforce the utility of genetic liability for EA as a valuable tool for understanding the genetic interplay between dyslexia and academic outcomes. ETHICAL INFORMATION The project was approved by the Ethics Committee at the Hospital Universitari Vall d’Hebron (PR(AG)491/2022) and the CEIm Fundació Sant Joan de Déu (PIC-154-22) and written informed consent was obtained from parents or caregivers. Declarations DECLARATION OF INTERESTS J.A.R.Q was on the speakers’ bureau and/or acted as consultant for Biogen, Janssen-Cilag, Novartis, Shire, Takeda, Bial, Shionogi, Sincrolab, Novartis, BMS, Medice, Rubió, Uriach, Technofarma and Raffo in the last 3 years. He also received travel awards (air tickets + hotel) for taking part in psychiatric meetings from Janssen-Cilag, Rubió, Shire, Takeda, Shionogi, Bial and Medice. The Department of Psychiatry chaired by him received unrestricted educational and research support from the following companies in the last 3 years: Janssen- Cilag, Shire, Oryzon, Roche, Psious, and Rubió. The rest of authors have nothing to disclose. FUNDING This work was supported by the Agència de Gestió d’Ajuts Universitaris i de Recerca (AGAUR, 2017SGR-1461, 2021SGR-00840, 2021SGR-01093); the Instituto de Salud Carlos III (PI20/00041, PI22/00464, PI23/00404, PI23/00026, PI24/00195, CP22/00128 to M.S.A, CP22/00026 to S.A, FI23/00152 to P.C.G); the Network Center for Biomedical Research (CIBER) to J.C.D and U.Z.A.; the European Regional Development Fund (ERDF); the ECNP Network ‘ADHD across the Lifespan’; “la Marató de TV3” (202228-30 and 202228-31); the European Union H2020 Programme (H2020/2014–2020) under grant agreements no. 848228 (DISCOvERIE) and no. 2020604 (TIMESPAN); “Fundació ‘la Caixa’, Diputació de Barcelona, Pla Estratègic de Recerca i Innovació en Salut” (PERISSLT006/17/285); “Fundació Privada d'Investigació Sant Pau” (FISP); Ministry of Health of Generalitat de Catalunya; grant RYC2021-033573-I funded by MICIU/AEI/ 10.13039/501100011033 and European Union NextGenerationEU/PRTR to MM. ACKNOWLEDGMENTS The authors are grateful to families, students, and staff of the public primary schools (i.e., Joan Maragall, María Bores, Marqués de la Pobla, Martinet, Pins del Vallès, Puiggracios, Sant Jordi, Ramon Llull, Rivo Rubeo, Tagamanent and Teresa Berguedà), public secondary schools (i.e., Angeleta Ferrer i Sensat, Antoni Pous i Argila, Cal Gravat, Duc de Montblanc, Institut del Ter, Jaume Callís, Lacetània, Lluís de Peguera, Molí de la Vila, Montsuar, Pius Font i Quer, Vallbona d’Anoia, and Vil la Romana), and private schools (i.e., Airina, L'Ave Maria, Casals – Gràcia, Episcopal Lleida, La Farga, FEDAC Manresa, FEDAC Vic, Garbí Pere Vergés Esplugues, Institucio Igualada, Joviat, Oms i de Prat, Pies Mataró, Pureza de Maria, Regina Carmeli, Sagrats Cors Centelles, La Salle Manlleu, La Salle Manresa, Sant Miquel dels Sants, Thau Barcelona and Vedruna Escorial Vic) who kindly contribute in this research. We would like to thank the research participants and employees of 23andMe, Inc. for making this work possible. The genotyping service was carried out at the Genotyping Unit-CEGEN in the Spanish National Cancer Research Centre (CNIO), supported by Instituto de Salud Carlos III (ISCIII), Ministerio de Ciencia e Innovación. CEGEN is part of the initiative IMPaCTGENóMICA (IMP/00009) cofunded by ISCIII and the European Regional Development Fund (ERDF). DATA AVAILABILITY Data from the INSchool cohort included in this article is not publicly available due to limitations in ethical approvals and the summary data will be available upon reasonable request. The full GWAS summary statistics from the original 23andMe discovery studies have been made available through 23andMe to qualified researchers under agreements with 23andMe that protects participants privacy. Datasets will be made available at no cost for academic use, visit https://research.23andme.com/collaborate/#dataset-access/ for more information and to apply to access. All participants from 23andMe provided informed consent and volunteered to participate in the research online, under a protocol approved by the external AAHRPP-accredited insttutional review board, Ethical and Independent Review Services. As of 2022, Ethical and Independent Review Services is part of Salus institutional review board ( https://www.versiticlinicaltrials.org/salusirb ). References Erbeli F, Rice M, Paracchini S. Insights into Dyslexia Genetics Research from the Last Two Decades. Brain Sci 2021; 12: 27. Wagner RK, Zirps FA, Edwards AA, Wood SG, Joyner RE, Becker BJ et al. The Prevalence of Dyslexia: A New Approach to Its Estimation. J Learn Disabil 2020; 53: 354–365. Wilmot A, Pizzey H, Leitão S, Hasking P, Boyes M. Growing up with dyslexia: Child and parent perspectives on school struggles, self-esteem, and mental health. Dyslexia 2023; 29: 40–54. Ciulkinyte A, Mountford HS, Fontanillas P, Team R, Bates TC, Martin NG et al. Genetic neurodevelopmental clustering and dyslexia. medRxiv 2023. doi: 10.1101/2023.10.04.23296530 . Haworth CMA, Kovas Y, Harlaar N, Hayiou-Thomas ME, Petrill SA, Dale PS et al. Generalist genes and learning disabilities: A multivariate genetic analysis of low performance in reading, mathematics, language and general cognitive ability in a sample of 8000 12-year-old twins. J Child Psychol Psychiatry 2009; 50: 1318–1325. Doust C, Fontanillas P, Eising E, Gordon SD, Wang Z, Alagöz G et al. Discovery of 42 genome-wide significant loci associated with dyslexia. Nat Genet 2022; 54: 1621–1629. Verhoef E, Demontis D, Burgess S, Shapland CY, Dale PS, Okbay A et al. Disentangling polygenic associations between attention-deficit/hyperactivity disorder, educational attainment, literacy and language. Transl Psychiatry 2019; 9. doi: 10.1038/s41398-018-0324-2 . Luciano M, Hagenaars SP, Cox SR, Hill WD, Davies G, Harris SE et al. Single Nucleotide Polymorphisms Associated with Reading Ability Show Connection to Socio-Economic Outcomes. Behav Genet 2017; 47: 469–479. Richardson JTE, Wydell TN. The representation and attainment of students with dyslexia in UK higher education. 2003. Lee JJ, Wedow R, Okbay A, Kong E, Maghzian O, Zacher M et al. Gene discovery and polygenic prediction from a genome-wide association study of educational attainment in 1.1 million individuals. Nat Genet 2018; 50: 1112–1121. Branigan AR, Mccallum KJ, Freese J. Variation in the heritability of educational attainment: An international meta-analysis. Social Forces 2013; 92: 109–140. Cutler D, Lleras-Muney A. Education and Health: Evaluating Theories and Evidence. Cambridge, MA, 2006 doi: 10.3386/w12352 . Krapohl E, Rimfeld K, Shakeshaft NG, Trzaskowski M, McMillan A, Pingault JB et al. The high heritability of educational achievement reflects many genetically influenced traits, not just intelligence. Proc Natl Acad Sci U S A 2014; 111: 15273–15278. Ward ME, McMahon G, St Pourcain B, Evans DM, Rietveld CA, Benjamin DJ et al. Genetic variation associated with differential educational attainment in adults has anticipated associations with school performance in children. PLoS One 2014; 9. doi: 10.1371/journal.pone.0100248 . Leitão S, Dzidic P, Claessen M, Gordon J, Howard K, Nayton M et al. Exploring the impact of living with dyslexia: The perspectives of children and their parents. Int J Speech Lang Pathol 2017; 19: 322–334. Singer E. Coping with academic failure, a study of Dutch children with dyslexia. Dyslexia 2008; 14: 314–333. Carawan LW, Nalavany BA, Jenkins C. Emotional experience with dyslexia and self-esteem: The protective role of perceived family support in late adulthood. Aging Ment Health 2016; 20: 284–294. Humphrey N, Mullins PM. Self-concept and self‐esteem in developmental dyslexia. Journal of Research in Special Educational Needs 2002; 2. doi: 10.1111/j.1471-3802.2002.00163.x . Bosch R, Pagerols M, Rivas C, Sixto L, Bricollé L, Español-Martín G et al. Neurodevelopmental disorders among Spanish school-age children: prevalence and sociodemographic correlates. Psychol Med 2022; 52: 3062–3072. Cuetos F, Rodríguez B, Ruano E, Arribas D. PROLEC-R: Batería de evaluación de los procesos lectores. Revisada. . Madrid: TEA Ediciones., 2007. Cuetos F, Arribas D, Ramos JL. PROLEC-SE-R. Batería para la Evaluación de los Procesos Lectores en Secundaria y Bachillerato - Revisada. Madrid: TEA Ediciones, 2016 https://web.teaediciones.com/PROLEC-SE-R-Bateria-para-la-Evaluacion-de-los-Procesos-Lectores-en-Secundaria-y-Bachillerato—Revisada.aspx (accessed 2 Sep2024). Wechsler D. Wechsler intelligence scale for children–fifth edition (WISC-V). . San Antonio, TX: NCS Pearson, 2014. Bosch R, Pagerols M, Rivas C, Sixto L, Bricollé L, Español-Martín G et al. Neurodevelopmental disorders among Spanish school-age children: prevalence and sociodemographic correlates. Psychol Med 2022; 52: 3062–3072. Achenback TM, Rescorla LA. Manual for ASEBA school-age forms & profiles . University of Vermont, Research Center for Children, Youth & Families, 2001. Cabana-Domínguez J, Bosch R, Soler Artigas M, Alemany S, Llonga N, Vilar-Ribó L et al. Dissecting the polygenic contribution of attention-deficit/hyperactivity disorder and autism spectrum disorder on school performance by their relationship with educational attainment. Molecular Psychiatry 2024 2024;: 1–13. Das S, Forer L, Schönherr S, Sidore C, Locke AE, Kwong A et al. Next-generation genotype imputation service and methods. Nat Genet 2016; 48: 1284–1287. Choi SW, Mak TS-H, O’Reilly PF. Tutorial: a guide to performing polygenic risk score analyses. Nat Protoc 2020; 15: 2759–2772. Andreu-Bernabeu Á, Díaz-Caneja CM, Costas J, De Hoyos L, Stella C, Gurriarán X et al. Polygenic contribution to the relationship of loneliness and social isolation with schizophrenia. Nat Commun 2022; 13. doi: 10.1038/S41467-021-27598-6 . Bulik-Sullivan B, Loh PR, Finucane HK, Ripke S, Yang J, Patterson N et al. LD score regression distinguishes confounding from polygenicity in genome-wide association studies. Nat Genet 2015; 47: 291–295. Frei O, Holland D, Smeland OB, Shadrin AA, Fan CC, Maeland S et al. Bivariate causal mixture model quantifies polygenic overlap between complex traits beyond genetic correlation. Nat Commun 2019; 10. doi: 10.1038/s41467-019-10310-0 . Finucane HK, Bulik-Sullivan B, Gusev A, Trynka G, Reshef Y, Loh PR et al. Partitioning heritability by functional annotation using genome-wide association summary statistics. Nat Genet 2015; 47: 1228–1235. Christensen RHB. Regression Models for Ordinal Data [R package ordinal version 2023.12-4]. 2023. Fox J. Effect Displays in R for Generalised Linear Models. J Stat Softw 2003; 8. doi: 10.18637/jss.v008.i15 . Hollingshead AB. Four Factor Index of Social Status . Yale Journal of Sociology, 2011. Lu Q, Li B, Ou D, Erlendsdottir M, Powles RL, Jiang T et al. A Powerful Approach to Estimating Annotation-Stratified Genetic Covariance via GWAS Summary Statistics. Am J Hum Genet 2017; 101: 939–964. Bansal V, Mitjans M, Burik CAP, Linnér RK, Okbay A, Rietveld CA et al. Genome-wide association study results for educational attainment aid in identifying genetic heterogeneity of schizophrenia. Nature Communications 2018 9:1 2018; 9: 1–12. Gialluisi A, Andlauer TFM, Mirza-Schreiber N, Moll K, Becker J, Hoffmann P et al. Genome-wide association study reveals new insights into the heritability and genetic correlates of developmental dyslexia. Mol Psychiatry 2021; 26: 3004–3017. Dahle AE, Knivsberg AM. Internalizing, externalizing and attention problems in dyslexia. Scandinavian Journal of Disability Research 2014; 16: 179–193. Livingston EM, Siegel LS, Ribary U. Developmental dyslexia: emotional impact and consequences. Aust J Learn Diffic 2018; 23: 107–135. Weckström T, Elovainio M, Pulkki-Råback L, Suokas K, Komulainen K, Mullola S et al. School achievement in adolescence and the risk of mental disorders in early adulthood: a Finnish nationwide register study. Mol Psychiatry 2023; 28: 3104–3110. Novita S. Secondary symptoms of dyslexia: a comparison of self-esteem and anxiety profiles of children with and without dyslexia. Eur J Spec Needs Educ 2016; 31: 279–288. Brittain H, Vaillancourt T. Longitudinal associations between academic achievement and depressive symptoms in adolescence: Methodological considerations and analytical approaches for identifying temporal priority. In: Advances in Child Development and Behavior . Academic Press Inc., 2023, pp 327–355. Howe LJ, Nivard MG, Morris TT, Hansen AF, Rasheed H, Cho Y et al. Within-sibship genome-wide association analyses decrease bias in estimates of direct genetic effects. Nat Genet 2022; 54: 581–592. Nivard MG, Belsky DW, Harden KP, Baier T, Andreassen OA, Ystrøm E et al. More than nature and nurture, indirect genetic effects on children’s academic achievement are consequences of dynastic social processes. Nat Hum Behav 2024; 8: 771–778. Hugh-Jones D, Verweij KJH, St. Pourcain B, Abdellaoui A. Assortative mating on educational attainment leads to genetic spousal resemblance for polygenic scores. Intelligence 2016; 59: 103–108. Geredakis A, Vergou M, Zakopoulou V. Complementary and Alternative Approaches to Therapeutic Dyslexia Intervention. World J Res Rev 2017; 4: 36–50. Duff DM, Hendricks AE, Fitton L, Adlof SM. Reading and Math Achievement in Children With Dyslexia, Developmental Language Disorder, or Typical Development: Achievement Gaps Persist From Second Through Fourth Grades. J Learn Disabil 2023; 56: 371–391. Ashkenazi S, Black JM, Abrams DA, Hoeft F, Menon V. Neurobiological Underpinnings of Math and Reading Learning Disabilities. J Learn Disabil 2013; 46: 549–569. Additional Declarations Yes J.A.R.Q was on the speakers’ bureau and/or acted as consultant for Biogen, Janssen-Cilag, Novartis, Shire, Takeda, Bial, Shionogi, Sincrolab, Novartis, BMS, Medice, Rubió, Uriach, Technofarma and Raffo in the last 3 years. He also received travel awards (air tickets + hotel) for taking part in psychiatric meetings from Janssen-Cilag, Rubió, Shire, Takeda, Shionogi, Bial and Medice. The Department of Psychiatry chaired by him received unrestricted educational and research support from the following companies in the last 3 years: Janssen- Cilag, Shire, Oryzon, Roche, Psious, and Rubió. The rest of authors have nothing to disclose. Supplementary Files Supplementarytables.xlsx Supplementary tables SupplementaryFigures.docx Supplementary figures Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: revise 13 Jan, 2026 Review # 2 received at journal 22 Dec, 2025 Reviewer # 2 agreed at journal 10 Dec, 2025 Review # 1 received at journal 13 Aug, 2025 Reviewer # 1 agreed at journal 12 Aug, 2025 Reviewers invited by journal 11 Aug, 2025 Editor assigned by journal 28 Jul, 2025 Submission checks completed at journal 28 Jul, 2025 First submitted to journal 25 Jul, 2025 Unknown event 25 Jul, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7206327","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":498971761,"identity":"7c680297-fb50-4cc9-8d0a-ba2d1e02ba81","order_by":0,"name":"Judit Cabana-Domínguez","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4klEQVRIiWNgGAWjYBACxgYGgwMghgGI+MBwgIc4LQegWhhnEKMFrBimhZkHzCQAmGckbzz8gcFO3pz97MHPtm13ZMwbmI99/ILPYTPSCoAOSzbc2ZOXLJ3b9oxH5gBb8mwZvFpyQH45wLjhQI6BdM6ZwzwSDDzGzBJEaLHfcP6N8W8LUrQkbriRYybNUAHRwvgBn5aeZwUHzhgkJ++c8cbMsgekhZktmRmPDgbD9uTNHyoq7Gy38+cY3/hhcNhegr35MOMPfFoaQKQBshAzOIJwA3nszsVnyygYBaNgFIw4AADqck78PfzZ3wAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-4732-7284","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":true,"prefix":"","firstName":"Judit","middleName":"","lastName":"Cabana-Domínguez","suffix":""},{"id":498971762,"identity":"182e6f99-1077-48f3-ab8d-5c0101e8d85f","order_by":1,"name":"Marta Graell","email":"","orcid":"","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":false,"prefix":"","firstName":"Marta","middleName":"","lastName":"Graell","suffix":""},{"id":498971763,"identity":"42484e08-767a-40ff-8aa8-e5dcd225a5ad","order_by":2,"name":"Rosa Bosch","email":"","orcid":"","institution":"Vall d'Hebron Research Institute","correspondingAuthor":false,"prefix":"","firstName":"Rosa","middleName":"","lastName":"Bosch","suffix":""},{"id":498971764,"identity":"f16d4b80-8b55-45b5-9d4b-ab4d40980885","order_by":3,"name":"María Soler Artigas","email":"","orcid":"https://orcid.org/0000-0002-3213-1107","institution":"Vall d'Hebron Research Institute","correspondingAuthor":false,"prefix":"","firstName":"María","middleName":"Soler","lastName":"Artigas","suffix":""},{"id":498971765,"identity":"2fc5a3d9-c407-4884-8702-02f5a0c4bc88","order_by":4,"name":"Silvia Alemany","email":"","orcid":"https://orcid.org/0000-0002-7925-6767","institution":"vhir","correspondingAuthor":false,"prefix":"","firstName":"Silvia","middleName":"","lastName":"Alemany","suffix":""},{"id":498971766,"identity":"e5a9c078-1c8c-412d-ae2a-1bfe9eccf805","order_by":5,"name":"Natalia Llonga","email":"","orcid":"https://orcid.org/0000-0003-4325-6201","institution":"Vall d'Hebron Research Institute (VHIR), Universitat Autònoma de Barcelona","correspondingAuthor":false,"prefix":"","firstName":"Natalia","middleName":"","lastName":"Llonga","suffix":""},{"id":498971767,"identity":"a0347797-9e1f-4c1a-86d0-d5135097abff","order_by":6,"name":"Pau Carabí-Gassol","email":"","orcid":"","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":false,"prefix":"","firstName":"Pau","middleName":"","lastName":"Carabí-Gassol","suffix":""},{"id":498971768,"identity":"3ca428c1-38bb-4f8e-9833-69177bcb32a6","order_by":7,"name":"Uxue Zubizarreta-Arruti","email":"","orcid":"","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":false,"prefix":"","firstName":"Uxue","middleName":"","lastName":"Zubizarreta-Arruti","suffix":""},{"id":498971769,"identity":"22cfae19-eae7-46bc-bcda-10c3744d528d","order_by":8,"name":"Valeria Macias-Chimborazo","email":"","orcid":"","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":false,"prefix":"","firstName":"Valeria","middleName":"","lastName":"Macias-Chimborazo","suffix":""},{"id":498971770,"identity":"a4e17cf9-5a63-4ec8-a182-d577e7cd93f9","order_by":9,"name":"Marina Mitjans","email":"","orcid":"https://orcid.org/0000-0001-5543-7320","institution":"Max Planck Institute of Experimental Medicine, Göttingen, Germany","correspondingAuthor":false,"prefix":"","firstName":"Marina","middleName":"","lastName":"Mitjans","suffix":""},{"id":498971771,"identity":"6f49680b-bd63-45ac-8424-023d9af5edfe","order_by":10,"name":"Mireia Pagerols","email":"","orcid":"","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":false,"prefix":"","firstName":"Mireia","middleName":"","lastName":"Pagerols","suffix":""},{"id":498971772,"identity":"fb43cd5d-2023-4f2c-822a-fc9df57a6f49","order_by":11,"name":"Èlia Pagespetit","email":"","orcid":"","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":false,"prefix":"","firstName":"Èlia","middleName":"","lastName":"Pagespetit","suffix":""},{"id":498971773,"identity":"d90daba0-f2e2-49e4-b077-4a7afeb8bc7c","order_by":12,"name":"Raquel Prat","email":"","orcid":"","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":false,"prefix":"","firstName":"Raquel","middleName":"","lastName":"Prat","suffix":""},{"id":498971774,"identity":"c3c8bd25-a4b4-481a-90c2-0eb0e09900db","order_by":13,"name":"Júlia Puigbó","email":"","orcid":"","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":false,"prefix":"","firstName":"Júlia","middleName":"","lastName":"Puigbó","suffix":""},{"id":498971775,"identity":"23e4fada-6d0c-40e3-88b0-ddf0a8d0ea98","order_by":14,"name":"Cristina Rivas","email":"","orcid":"","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":false,"prefix":"","firstName":"Cristina","middleName":"","lastName":"Rivas","suffix":""},{"id":498971776,"identity":"c84423fa-7386-4814-8411-4091bab31f19","order_by":15,"name":"Laura Martinez","email":"","orcid":"","institution":"Vall d'Hebron Research Institute (VHIR)","correspondingAuthor":false,"prefix":"","firstName":"Laura","middleName":"","lastName":"Martinez","suffix":""},{"id":498971777,"identity":"2611833a-089f-42af-863a-19c9ec0058df","order_by":16,"name":"Josep Antoni Ramos-Quiroga","email":"","orcid":"https://orcid.org/0000-0003-1622-0350","institution":"Hospital Universitari Vall d'Hebron.","correspondingAuthor":false,"prefix":"","firstName":"Josep","middleName":"Antoni","lastName":"Ramos-Quiroga","suffix":""},{"id":498971778,"identity":"671288c8-1a8a-4e16-b3b7-59ebd5ad3876","order_by":17,"name":"Miguel Casas","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Miguel","middleName":"","lastName":"Casas","suffix":""},{"id":498971779,"identity":"435ed62f-f442-40e6-963b-328d88fc046a","order_by":18,"name":"Marta Ribasés","email":"","orcid":"https://orcid.org/0000-0003-1039-1116","institution":"Vall d'Hebron Research Institute (VHIR), Universitat Autònoma de Barcelona","correspondingAuthor":false,"prefix":"","firstName":"Marta","middleName":"","lastName":"Ribasés","suffix":""}],"badges":[],"createdAt":"2025-07-24 13:57:47","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7206327/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7206327/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":89456451,"identity":"416a7b98-a70e-4034-b619-d88c6847327e","added_by":"auto","created_at":"2025-08-20 07:06:47","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":241328,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eResults from genetic overlap between dyslexia (DYS) and educational attainment (EA).\u003c/strong\u003e A) Venn diagram depicting the estimated number of trait-influencing variants shared (gray) between dyslexia and educational attainment. Unique variants for each trait are depicted in blue for educational attainment and orange for dyslexia. The number of trait-influencing variants in thousands is shown, with the standard error in thousands provided in parentheses. The size of the circles reflects the polygenicity of each phenotype, with larger circles corresponding to greater polygenicity. The estimated genetic correlation (r\u003csub\u003eg\u003c/sub\u003e) is shown in the bar. The blue color indicates negative genetic correlation. B) Log-likelihood curves highlighting the goodness of model fit. The minimum point indicates the best-fitting model estimate of the number of influencing variants shared between two traits. C) and D) depict conditional Q–Q plots of observed versus expected −log10 P-values in the primary trait as a function of significance of association with a secondary trait at the level of p ≤ 0.1 (orange lines), p ≤ 0.01 (green lines), p ≤ 0.001 (red lines). The blue line indicates all SNPs. Dotted lines in blue, orange, green, and red indicate model predictions for each stratum. Black dotted line is the expected Q–Q plot under null (no SNPs associated with the phenotype). For this analysis, summary statistics from GWAS meta-analysis on EA (N=1,131,881 individuals) \u003csup\u003e10\u003c/sup\u003e and dyslexia (N=51,800 cases and 1,087,070 controls, N\u003csub\u003eeffective\u003c/sub\u003e=197,776 individuals) \u003csup\u003e6\u003c/sup\u003e were used.\u003c/p\u003e","description":"","filename":"Figure171.png","url":"https://assets-eu.researchsquare.com/files/rs-7206327/v1/32c98b339a7e85fd623592f1.png"},{"id":89453664,"identity":"1dbda41c-0300-4e98-971c-d109438d3b85","added_by":"auto","created_at":"2025-08-20 06:42:47","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":112170,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eContribution of the genome-wide polygenic score of educational attainment (PGS\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003eEA\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003e)\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003e \u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003eand dyslexia (PGS\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003eDYS\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003e) dissected by their relationship with EA to school performance.\u003c/strong\u003e PGS\u003csub\u003eDYS\u003c/sub\u003e were constructed considering four subsets of SNPs in our in-house cohort of 4,274 school-aged participants: variants not associated with EA (P\u003csub\u003eEA\u003c/sub\u003e\u0026gt;0.05; PGS\u003csub\u003eDYS_noEA\u003c/sub\u003e), variants associated with EA (P\u003csub\u003eEA\u003c/sub\u003e £ 0.05; PGS\u003csub\u003eDYS_EA\u003c/sub\u003e), variants showing concordant (PGS\u003csub\u003eDYSconcordant\u003c/sub\u003e) and discordant (PGS\u003csub\u003eDYSdiscordant\u003c/sub\u003e) direction of the effect in EA and DYS. A) Percentage of variance (Nagelkerke’s pseudo-R\u003csup\u003e2\u003c/sup\u003e) explained by PGS\u003csub\u003eEA, \u003c/sub\u003ePGS\u003csub\u003eDYS\u003c/sub\u003e and PGS\u003csub\u003eDYS\u003c/sub\u003e according to the four partitions described above. B) Quantile plots of odds ratios (OR) with 95% confidence intervals for PGS\u003csub\u003eEA, \u003c/sub\u003ePGS\u003csub\u003eDYS\u003c/sub\u003e and PGS\u003csub\u003eDYS\u003c/sub\u003e according to the four partitions described above. The target sample was divided into quintiles and school performance of each quintile was compared to the first quintile using ordinal mixed-effect models with age, sex, socio-economic status (SES) and 20 principal genetic components as fixed effects. Significant comparisons (P\u003csub\u003eFDR\u003c/sub\u003e \u0026lt; 0.05) are indicated with an asterisk.\u003c/p\u003e","description":"","filename":"Figure172.png","url":"https://assets-eu.researchsquare.com/files/rs-7206327/v1/102dc922b52e306f4074a692.png"},{"id":89455057,"identity":"d43a7541-6d30-46f1-a96d-03b65c41e731","added_by":"auto","created_at":"2025-08-20 06:50:47","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":37754,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDissection of childhood psychopathology by genome-wide polygenic scores (PGS) in our in-house cohort of 4,274 school-aged participants.\u003c/strong\u003e A) Effect of association analyses for CBCL subscales (y axis) and PGS score for educational attainment (EA), dyslexia (DYS) and dyslexia dissected by their relationship with EA (x axis). The color gradient indicates the strength and direction of the association, with positive associations being shown in pink and negative associations in blue. B) Z-scores from association analyses for school performance in the three subjects and CBCL subscales and PGS\u003csub\u003eDYSdiscordant\u003c/sub\u003e (x axis) and PGS\u003csub\u003eDYSconcordant\u003c/sub\u003e (y axis). Point size corresponds to -log10(P-value) with concordant PGS in pink and discordant PGS in blue. Red dotted lines indicate significant results after FDR correction.\u003c/p\u003e","description":"","filename":"Figure173.png","url":"https://assets-eu.researchsquare.com/files/rs-7206327/v1/dc3f4b584fa4376579e17422.png"},{"id":89453666,"identity":"605ab78b-d266-4a22-8294-c353bdb89f1e","added_by":"auto","created_at":"2025-08-20 06:42:47","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":105587,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAnnotation-stratified genetic covariance between dyslexia and related disorders and traits.\u003c/strong\u003e Genetic covariances for five subsets of SNPs: all SNPs from EA (dark blue), all SNPs from dyslexia (light blue), variants not associated with EA (green) and variants associated with EA with concordant (orange) and discordant (pink) direction of the effect in DYS and EA. FDR-corrected significant associations (\u003cem\u003eP\u003c/em\u003e\u003csub\u003e\u003cem\u003eFDR\u003c/em\u003e\u003c/sub\u003e \u0026lt; 0.05) are marked with an asterisk.\u003c/p\u003e","description":"","filename":"Figure174.png","url":"https://assets-eu.researchsquare.com/files/rs-7206327/v1/6740f7a7946f2d00b48cf60f.png"},{"id":89457431,"identity":"31e747d5-7805-4865-b8eb-d6b0cea1cfec","added_by":"auto","created_at":"2025-08-20 07:14:48","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1673070,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7206327/v1/66188bee-0387-4245-9087-ef10bfd499eb.pdf"},{"id":89456090,"identity":"4a643115-b98b-4f32-8d13-2ec701c2c6dd","added_by":"auto","created_at":"2025-08-20 06:58:47","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":59054,"visible":true,"origin":"","legend":"Supplementary tables","description":"","filename":"Supplementarytables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7206327/v1/cce37684ffb88f0acc31140d.xlsx"},{"id":89453670,"identity":"74cb43d8-b469-415c-94a1-7ac4335256ad","added_by":"auto","created_at":"2025-08-20 06:42:47","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":1781077,"visible":true,"origin":"","legend":"Supplementary figures","description":"","filename":"SupplementaryFigures.docx","url":"https://assets-eu.researchsquare.com/files/rs-7206327/v1/9ffb4d176a5cda577600284a.docx"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e\nJ.A.R.Q was on the speakers’ bureau and/or acted as consultant for Biogen, Janssen-Cilag, Novartis, Shire, Takeda, Bial, Shionogi, Sincrolab, Novartis, BMS, Medice, Rubió, Uriach, Technofarma and Raffo in the last 3 years. He also received travel awards (air tickets + hotel) for taking part in psychiatric meetings from Janssen-Cilag, Rubió, Shire, Takeda, Shionogi, Bial and Medice. The Department of Psychiatry chaired by him received unrestricted educational and research support from the following companies in the last 3 years: Janssen- Cilag, Shire, Oryzon, Roche, Psious, and Rubió. The rest of authors have nothing to disclose.","formattedTitle":"\u003cp\u003eTowards Disentangling the Polygenic Contribution of Dyslexia to School Performance\u003c/p\u003e","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eDyslexia is a neurodevelopmental disorder commonly diagnosed in school-aged children, characterized by a persistent impairment in specific learning skills such as reading accuracy, spelling, grammar, or reading fluency. It is found worldwide and across all languages, including different alphabetic orthographies and logographic languages \u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e, with prevalence estimates ranging from 5\u0026ndash;20% \u003csup\u003e2\u003c/sup\u003e. Dyslexia typically manifests in educational settings, but its impacts often extends beyond the classroom, affecting self-esteem, social interactions, and long-term occupational prospects \u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. Additionally, dyslexia frequently co-occurs with other psychiatric and neurodevelopmental disorders, including attention-deficit/hyperactivity disorder (ADHD) and language disorders \u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eDyslexia is a complex disorder with an estimated heritability of up to 70% from twin and family studies \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. The largest genome-wide association study (GWAS) on dyslexia to date (51,800 cases and 1,087,070 controls) identified 42 independent hits and significant genetic correlations between dyslexia and 63 phenotypes including several mental disorders and behavioral traits (e.g., ADHD and risk-taking behavior) as well as measures of academic performance (e.g., educational attainment (EA) and age completed education) \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eDyslexia is related with low EA \u003csup\u003e\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e, which represents the highest level of education an individual has completed. EA is commonly used as proxy to quantify educational performance and is also linked to different developmental disorders, as well as social, economic and health outcomes \u003csup\u003e\u003cspan additionalcitationids=\"CR11\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Similar to dyslexia, EA is a highly polygenic trait, with 1,271 independent hits identified in one of the largest GWAS published to date in about 1.1\u0026nbsp;million individuals \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Dyslexia is also negatively associated with school performance, which is considered an early-life intermediate phenotype that may anticipate EA \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. Children with dyslexia have an increased risk of obtaining lower grades compared to other peers and, consequently, a greater risk of falling behind on, or dropping out from, school \u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. These academic challenges can result in a negative feedback loop, worsening the self-confidence and motivation of these students, further aggravating their academic difficulties \u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Furthermore, the stigma around learning disabilities can lead to social isolation, contributing to a negative school experience \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eDyslexia and EA are complex and polygenic traits, but their genetic relationship with school performance remains underexplored. We aimed to clarify whether the association between the genetic liability of dyslexia with school performance is partially explained by the genetic liability for EA. Using a deeply phenotyped cohort of 4,274 school-aged children, we sought to: (i) assess the shared genetic architecture between dyslexia and EA; (ii) explore the association of genetic liability of dyslexia and EA with school performance; (iii) determine whether the genetic contribution of dyslexia to school performance is influenced by the genetic liability of EA; and (iv) evaluate the association of dyslexia with childhood psychopathology and other psychiatric disorders and related traits, as well as assess the role of EA in these associations (Supplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e).\u003c/p\u003e"},{"header":"MATERIALS AND METHODS","content":"\u003cp\u003e\u003cb\u003eStudy sample, school performance and clinical assessment\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe study sample from the INSchool cohort included 4,274 school-aged participants (mean age\u0026thinsp;=\u0026thinsp;10.03 years, SD\u0026thinsp;=\u0026thinsp;2.95, ranging from 5 to 17 years; 43.7% females), representative of the general school-aged population, recruited in 45 primary and secondary schools across Catalonia, Spain \u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. The inclusion criteria for participants were European ancestry and written informed consent from parents or caregivers. School performance data was provided by schools and assessed with school grades in three subjects: mathematics, primary language and foreign language. Grades were described on a four-point scale, being A: excellent performance, B: good performance, C: adequate performance and D: underperformance.\u003c/p\u003e\u003cp\u003eDyslexia was diagnosed in 9.66% of the 4,274 participants (n\u0026thinsp;=\u0026thinsp;413). Assessment was conducted using the Battery for the Evaluation of Reading Processes, Revised (PROLEC-R) \u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e and the Battery for the Evaluation of Reading Processes in Junior and Senior High-School Students, Revised (PROLEC-SE-R) \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. Participants with intellectual disability were excluded based on results from the Wechsler Intelligence Scale for Children (WISC) \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e, as previously described \u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eChildhood psychopathology was assessed using the Child Behavior Checklist for ages 6\u0026ndash;18 (CBCL-6/18) from Achenbach System of Empirically Based Assessment (ASEBA) \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e completed by parents or surrogates in most of the school-aged sample (96%; n\u0026thinsp;=\u0026thinsp;4,084).\u003c/p\u003e\u003cp\u003e\u003cb\u003eGenotyping, imputation and quality control\u003c/b\u003e\u003c/p\u003e\u003cp\u003eGenomic DNA isolated from saliva samples or buccal swabs was genotyped in three different waves with the Illumina Infinium PsychChip_v1.0 array for wave 1 (n\u0026thinsp;=\u0026thinsp;793) or the Infinium Global Screening Array-24 version_2 (GSA_v2) for waves 2 (n\u0026thinsp;=\u0026thinsp;2732) and 3 (n\u0026thinsp;=\u0026thinsp;749) (Illumina, CA, San Diego, USA). Pre-imputation quality control was done with PLINK_2.0 software as described in Cabana-Dom\u0026iacute;nguez et al. 2024 \u003csup\u003e25\u003c/sup\u003e, prepared for imputation using McCarthy tools, and imputed with the Michigan Imputation Server separately \u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. After imputation, we only kept SNPs with an imputation INFO score\u0026thinsp;\u0026gt;\u0026thinsp;0.8, minor allele frequency (MAF)\u0026thinsp;\u0026gt;\u0026thinsp;0.01 and SNP call rate\u0026thinsp;\u0026gt;\u0026thinsp;0.95. Then, samples from the three GWAS waves were merged and only SNPs present in all of them, with MAF\u0026thinsp;\u0026gt;\u0026thinsp;0.01, SNP call rate\u0026thinsp;\u0026gt;\u0026thinsp;0.95 and differential case/control missingness rate\u0026thinsp;\u0026lt;\u0026thinsp;0.2 in the overall sample were considered for subsequent analyses, ending up with 3,725,956 SNPs.\u003c/p\u003e\u003cp\u003e\u003cb\u003eGWAS summary statistics\u003c/b\u003e\u003c/p\u003e\u003cp\u003eSummary statistics from GWAS meta-analysis on EA (N\u0026thinsp;=\u0026thinsp;1,131,881 individuals) \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e and dyslexia (N\u0026thinsp;=\u0026thinsp;51,800 cases and 1,087,070 controls, N\u003csub\u003eeffective\u003c/sub\u003e=197,776 individuals) \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e were obtained from 23andMe, Inc. The effective sample size for dyslexia was calculated as N\u003csub\u003eeffective\u003c/sub\u003e= 4/(1/N\u003csub\u003ecases\u003c/sub\u003e + 1/N\u003csub\u003econtrols\u003c/sub\u003e). SNPs with minor allele frequency (MAF)\u0026thinsp;\u0026lt;\u0026thinsp;0.01 or SNPs located on the X, Y and mitochondrial chromosomes were filtered out.\u003c/p\u003e\u003cp\u003e\u003cb\u003eGenome-wide polygenic scores (PGS)\u003c/b\u003e\u003c/p\u003e\u003cp\u003ePGS were constructed in our in-house sample of 4,274 school-aged participants using summary statistics of EA and dyslexia with the PRS-CS \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e and PLINK_2.0 softwares. Scores were calculated using the common set of SNPs between EA and dyslexia (N\u003csub\u003ecommonSNP\u003c/sub\u003e=6,429,800). Once computed, all PGS were standardized to a mean of 0 and a standard deviation of 1. Additionally, PGS for dyslexia (PGS\u003csub\u003eDYS\u003c/sub\u003e) were constructed with four subsets of SNPs based of their contribution to EA, as previously described \u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e,\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. Prior to subsetting, GWAS summary statistics were harmonized to ensure alignment of effect alleles across both studies. Then, genetic liability of dyslexia was divided in hierarchical subsets as follows (Supplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e): (i) variants not associated with EA (PGS\u003csub\u003eDYS_noEA\u003c/sub\u003e; P\u003csub\u003eEA\u003c/sub\u003e\u0026gt;0.05; N\u003csub\u003eSNPs\u003c/sub\u003e = 5,004,441), and (ii) variants associated with EA (PGS\u003csub\u003eDYS_EA\u003c/sub\u003e; P\u003csub\u003eEA\u003c/sub\u003e \u0026le;0.05; N\u003csub\u003eSNPs\u003c/sub\u003e = 1,425,359). The second group was divided in two additional subsets based on the direction of effects in dyslexia and EA: (iii) variants showing consistent direction of effect in both dyslexia and EA (PGS\u003csub\u003eDYSconcordant\u003c/sub\u003e; Beta\u003csub\u003eDYS\u003c/sub\u003e\u0026gt;0 and Beta\u003csub\u003eEA\u003c/sub\u003e\u0026gt;0 or Beta\u003csub\u003eDYS\u003c/sub\u003e\u0026lt;0 and Beta\u003csub\u003eEA\u003c/sub\u003e\u0026lt;0; N\u003csub\u003eSNPs\u003c/sub\u003e = 636,190), and (iv) variants showing opposite direction of effect in both dyslexia and EA (PGS\u003csub\u003eDYSdiscordant\u003c/sub\u003e; Beta\u003csub\u003eDYS\u003c/sub\u003e\u0026gt;0 and Beta\u003csub\u003eEA\u003c/sub\u003e\u0026lt;0 or Beta\u003csub\u003eDYS\u003c/sub\u003e\u0026lt;0 and Beta\u003csub\u003eEA\u003c/sub\u003e\u0026gt;0; N\u003csub\u003eSNPs\u003c/sub\u003e = 789,169).\u003c/p\u003e\u003cp\u003e\u003cb\u003eStatistical analyses\u003c/b\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eSNP-based heritability and genetic overlap\u003c/b\u003e\u003c/p\u003e\u003cp\u003eSummary statistics from GWAS meta-analysis on EA \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e and dyslexia \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e were used to estimate the SNP-based heritability (h\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e\u003csub\u003eSNP\u003c/sub\u003e) and the genetic correlation between EA and dyslexia using LDSC_v1.0 \u003csup\u003e29\u003c/sup\u003e. All analyses were restricted to Hapmap3 SNPs and considering a population prevalence for dyslexia of 10% \u003csup\u003e6\u003c/sup\u003e. In addition, MiXeR (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/precimed/mixer\u003c/span\u003e\u003cspan address=\"https://github.com/precimed/mixer\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) \u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e was used to calculate the number of trait-influencing SNPs with the univariate model, and the genetic overlap between traits with the bivariate model. The Dice coefficient explains the proportion of shared SNPs between traits, and the Akaike Information Criterion (AIC) was used to determine the model fit (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/precimed/mixer#aic-bic-interpretation\u003c/span\u003e\u003cspan address=\"https://github.com/precimed/mixer#aic-bic-interpretation\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cb\u003ePartitioned heritability\u003c/b\u003e\u003c/p\u003e\u003cp\u003eSNP based partitioned heritability (h\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e\u003csub\u003eSNP\u003c/sub\u003e) was estimated for the three independent dyslexia partitioned subsets (DYS\u003csub\u003enoEA\u003c/sub\u003e, DYS\u003csub\u003econcordant\u003c/sub\u003e and DYS\u003csub\u003ediscordant\u003c/sub\u003e) and calculated using LDSC_v1.0 \u003csup\u003e31\u003c/sup\u003e. The heritability was computed for each partion using the GWAS summary statistics of dyslexia \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e and data from the 1000G as reference panel, and the analysis was restricted to Hapmap3 SNPs. Enrichment of genome-wide significant hits (P\u0026thinsp;\u0026lt;\u0026thinsp;5e-08) of dyslexia in the genome partitions described previously were calculated using a Chi-square test and Manhattan plots were obtained using the qqman R package.\u003c/p\u003e\u003cp\u003e\u003cb\u003eAssociation between PGS and school performance\u003c/b\u003e\u003c/p\u003e\u003cp\u003ePGS were associated with school performance using an ordinal mixed-effect model considering A as the highest and D as the lowest category using the ordinal R package \u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. The percentage of variance attributable to each PGS was calculated as the increase in Naggelkerke\u0026rsquo;s pseudo-r2 between models with and without the PGS. P-values were corrected for multiple comparisons via Bonferroni correction. To compare the effect of different PGS, the target sample was divided into five quintiles of increasing PGS and odds ratio were compared using the lowest quintile as reference and represented in quantile plots. The predicted probabilities of being in each category (A, B, C or D) were calculated for each subject in order to visualize how the PGS affects school performance and represented in probability plots using the effects R library \u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eSex-stratified analyses were conducted to examine whether the association between PGS and school performance differed between males and females. The same regression model described for the whole sample was followed. The regression model was modified to include the interaction term PGS*sex, allowing the effect of the PGS on school performance to vary by sex and capture any differential impact.\u003c/p\u003e\u003cp\u003eThe association between PGS and behavioral and emotional problems was tested using linear-mixed effects models. The CBCL-6/18 parent rating scales were used as continuous variables and square-root transformed because of skewness. P-values were corrected for multiple comparisons via the Benjamini-Hochberg False Discovery Rate (FDR) method (P\u003csub\u003eadj\u003c/sub\u003e\u0026lt;0.05).\u003c/p\u003e\u003cp\u003eAll analyses were adjusted for age, sex, socio-economic status (SES) and 20 genetic PC as fixed effects, as well as school as random effect to account for the multilevel nature of the data. The regression analyses for the quantile plots could not include school as a covariate, given the little variability for this variable in the smaller subsets of the data used for this analysis. SES was calculated using the Hollingshead Four-Factor Index based on parent\u0026rsquo;s education and occupation \u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003e\u003cb\u003eMediation analysis\u003c/b\u003e\u003c/p\u003e\u003cp\u003eWe investigated whether the effect of the PGS on school performance was mediated by the diagnosis of dyslexia in the INSchool cohort. The effect of PGS on the mediator was estimated using a logarithmic mixed-effect logistic regression model, and the effect of the mediator on school performance was estimated using an ordinal mixed-effect model. When these two models had significant associations (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05), the effect of the mediator and the direct effect of the PGS on school performance was estimated using ordinal mixed-effect model. All analyses were adjusted for age, sex, SES and 20 genetic PC as fixed effects, as well as school as random effect. Partial mediation was determined when the direct effect remained significant but attenuated.\u003c/p\u003e\u003cp\u003e\u003cb\u003eGenetic covariance\u003c/b\u003e\u003c/p\u003e\u003cp\u003eWe used annotation-stratified genetic covariance analysis to examine how three partitioned subsets of dyslexia-associated SNPs (DYS\u003csub\u003enoEA\u003c/sub\u003e, DYS\u003csub\u003econcordant\u003c/sub\u003e, and DYS\u003csub\u003ediscordant\u003c/sub\u003e) contribute to the genetic overlap between dyslexia and other traits. This analysis was conducted using the GeNetic cOVariance Analyzer (GNOVA) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/xtonyjiang/GNOVA\u003c/span\u003e\u003cspan address=\"https://github.com/xtonyjiang/GNOVA\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e, which estimates genetic covariance between two phenotypes while accounting for linkage disequilibrium (LD) and sample overlap. We included GWAS summary statistics for nine cognitive and language-related traits, seven psychiatric disorders, three substance use-related traits, and five traits related to well-being and socio-economic status (see Supplementary Table\u0026nbsp;1 for details). Genetic covariance was calculated using partial correlations, restricted to SNPs within each of the three dyslexia subsets, as well as across the full set of SNPs for dyslexia and educational attainment (EA). Following the developer\u0026rsquo;s recommendation, genomic covariance instead of correlation estimates were considered in all the analyses and P-values were adjusted for multiple comparisons using the FDR (P\u003csub\u003eadj\u003c/sub\u003e \u0026lt; 0.05).\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cp\u003e\u003cb\u003eGenetic architecture of dyslexia and EA\u003c/b\u003e\u003c/p\u003e\u003cp\u003eWe estimated a SNP-based heritability (h\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e\u003csub\u003eSNP\u003c/sub\u003e) of 0.179 for dyslexia (SE\u0026thinsp;=\u0026thinsp;0.0071) and 0.102 for EA (SE\u0026thinsp;=\u0026thinsp;0.0023) and found a negative genetic correlation between them (rg=-0.186, SE\u0026thinsp;=\u0026thinsp;0.019, P\u0026thinsp;=\u0026thinsp;1.75E-22) (Supplementary Table\u0026nbsp;2). MiXeR analysis revealed that both traits are highly polygenic, with 10,315 common variants for dyslexia and 14,403 for EA. Most of the variants that influence dyslexia are shared with EA (99.5%; n\u0026thinsp;=\u0026thinsp;10,214, SE\u0026thinsp;=\u0026thinsp;465, Dice coefficient\u0026thinsp;=\u0026thinsp;0.826), with a slightly higher proportion of variants with discordant direction of the effect between them (58.0%) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Supplementary Table\u0026nbsp;3). The genetic correlation was consistent with the stratified cross-phenotype Q-Q plots, where P-values for one trait are plotted conditioned on different association strengths for the other trait (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eContribution of the genetic liability for dyslexia and EA on school performance\u003c/b\u003e\u003c/p\u003e\u003cp\u003eGenome-wide polygenic scores for EA (PGS\u003csub\u003eEA\u003c/sub\u003e) were associated with higher odds of having a better performance in all three subjects studied in the 4,274 school-aged participants from the INSchool cohort (OR\u0026thinsp;\u0026gt;\u0026thinsp;1.38, P\u0026thinsp;\u0026lt;\u0026thinsp;4.41E-24). In contrast, PGS for dyslexia (PGS\u003csub\u003eDYS\u003c/sub\u003e) were associated with worse performance across all three subjects, being the effect greater in language-related subjects (first language (OR\u0026thinsp;=\u0026thinsp;0.79, 95% CI\u0026thinsp;=\u0026thinsp;0.74\u0026ndash;0.84) and foreign language (OR\u0026thinsp;=\u0026thinsp;0.80, 95% CI\u0026thinsp;=\u0026thinsp;0.75\u0026ndash;0.85)) than in mathematics (OR\u0026thinsp;=\u0026thinsp;0.89, 95% CI\u0026thinsp;=\u0026thinsp;0.84\u0026ndash;0.94). The proportion of variance in school performance explained by the genetic liability of EA ranges from 2.69\u0026ndash;3.88% across subjects, while the genetic liability of dyslexia accounts for 0.46\u0026ndash;1.65% (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eAssociation between school performance and genome-wide polygenic score for dyslexia (PGSDYS) based on its relationship with educational attainment (EA).\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSubject\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePGS\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eOR (95% CI)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eStatistic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eP-value\u003csup\u003e#\u003c/sup\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eR\u003csup\u003e2\u003c/sup\u003e*\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFirst Language\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.43 (1.34\u0026ndash;1.53)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e11.006\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e3.59E-28\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e3.23%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.79 (0.74\u0026ndash;0.84)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-7.873\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e3.45E-15\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e1.65%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS no EA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.84 (0.79\u0026ndash;0.89)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-5.887\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e3.93E-09\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.92%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS - EA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.81 (0.76\u0026ndash;0.86)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-7.141\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e9.29E-13\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e1.36%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS concordant\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.06 (1-1.13)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e2.020\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.044\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.11%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS discordant\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.73 (0.69\u0026ndash;0.78)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-9.984\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e1.79E-23\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e2.65%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eForeign Language\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.38 (1.30\u0026ndash;1.47)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e10.122\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e4.41E-24\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e2.69%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.80 (0.75\u0026ndash;0.85)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-7.637\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e2.23E-14\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e1.53%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS no EA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.85 (0.8\u0026ndash;0.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-5.621\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e1.90E-08\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.83%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS - EA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.82 (0.77\u0026ndash;0.86)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-6.952\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e3.61E-12\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e1.27%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS concordant\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.04 (0.98\u0026ndash;1.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1.356\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.175\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.05%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS discordant\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.76 (0.72\u0026ndash;0.81)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-9.085\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e1.04E-19\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e2.16%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMathematics\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.47 (1.38\u0026ndash;1.57)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e12.210\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e2.75E-34\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e3.88%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.89 (0.84\u0026ndash;0.94)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-4.191\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e2.78E-05\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.46%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS no EA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.94 (0.89-1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-2.026\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.043\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.11%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS - EA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.87 (0.82\u0026ndash;0.92)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-4.983\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e6.27E-07\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.65%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS concordant\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.10 (1.04\u0026ndash;1.16)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e3.185\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e1.45E-03\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.26%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDYS discordant\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.78 (0.74\u0026ndash;0.83)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-8.227\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cb\u003e1.92E-16\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e1.76%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"6\"\u003e\u003csup\u003e\u003cem\u003e#\u003c/em\u003e\u003c/sup\u003eP-values surpassing Bonferroni correction (P-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05/18\u0026thinsp;=\u0026thinsp;2.7E-03) are shown in bold. * Explained variance attributable to PGS calculated as the increase in Nagelkerke\u0026rsquo;s pseudo-R2 between an ordinal mixed-effect model with and without the PGS variable.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003ePGS comparisons across ranked quintiles showed that the odds for better performance in children at the highest quintile for PGS\u003csub\u003eEA\u003c/sub\u003e was, on average, more than twice that in children in the first quintile for the three subjects. In contrast, we found the opposite effect of PGS\u003csub\u003eDYS\u003c/sub\u003e, where children in the highest quantile for PGS\u003csub\u003eDYS\u003c/sub\u003e have worse performance than children in the first quantile (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and Supplementary Table\u0026nbsp;4). In mathematics, we found that only the subset of individuals with the highest PGS\u003csub\u003eDYS\u003c/sub\u003e (5th quantile) show significant differences from those in the 1st quantile. These results are consistent with the increased probability of reaching better performance (grades A or B) with higher PGS\u003csub\u003eEA\u003c/sub\u003e or lower PGS\u003csub\u003eDYS\u003c/sub\u003e, and worse performance (grades C or D) with higher PGS\u003csub\u003eDYS\u003c/sub\u003e or lower PGS\u003csub\u003eEA\u003c/sub\u003e (Supplementary Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cb\u003ePolygenic dissection of the contribution of dyslexia to school performance by their relationship with EA\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe genetic load of dyslexia was divided in variants not associated with EA (DYS\u003csub\u003enoEA\u003c/sub\u003e) and variants associated with EA (DYS\u003csub\u003eEA\u003c/sub\u003e), which were then divided into variants with a consistent effect (DYS\u003csub\u003econcordant\u003c/sub\u003e) and variants with opposing effect (DYS\u003csub\u003ediscordant\u003c/sub\u003e) (Supplementary Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Partitioned heritability estimates were calculated for the non-overlapping subsets of variants (DYS\u003csub\u003enoEA\u003c/sub\u003e, DYS\u003csub\u003ediscordant\u003c/sub\u003e and DYS\u003csub\u003econcordant\u003c/sub\u003e). We found that most of the h\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e\u003csub\u003eSNP\u003c/sub\u003e of dyslexia was distributed between DYS\u003csub\u003enoEA\u003c/sub\u003e and DYS\u003csub\u003ediscordant\u003c/sub\u003e, being the last one especially relevant considering that a small proportion of SNPs (16.8%) explains a considerable portion (29.1%) of the h\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e\u003csub\u003eSNP\u003c/sub\u003e of dyslexia (Supplementary Table\u0026nbsp;5A). In addition, genome-wide significant variants for dyslexia identified by Doust et al. \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e were enriched across all analyzed subsets (Supplementary Figure S3 and Table\u0026nbsp;5B).\u003c/p\u003e\u003cp\u003ePGS for dyslexia were calculated for the four subsets of genetic variants and tested for association with school performance in our in-house cohort of 4,274 school-aged children and adolescents. The results revealed similar patterns of association for performance in primary and foreign languages, but distinct patterns when comparing these language-related subjects with mathematics. PGS for dyslexia excluding EA-associated variants, PGS\u003csub\u003eDYS_noEA\u003c/sub\u003e, were associated with worse performance in first language (OR\u0026thinsp;=\u0026thinsp;0.84, 95% CI\u0026thinsp;=\u0026thinsp;0.79\u0026ndash;0.89) and foreign language (OR\u0026thinsp;=\u0026thinsp;0.85, 95% CI\u0026thinsp;=\u0026thinsp;0.8\u0026ndash;0.9), but showed no significant association with performance in mathematics. In contrast, PGS for dyslexia including EA-associated variants, PGS\u003csub\u003eDYS_EA\u003c/sub\u003e, were associated with worse performance across all three subjects (OR\u0026thinsp;\u0026lt;\u0026thinsp;0.94, adj-P\u0026thinsp;\u0026le;\u0026thinsp;0.046), although the proportion of variance explained was greater for language-related subjects (first language: 1,36% and foreign language: 1.27%) than for mathematics (0.65%) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). When dissecting variants shared with EA, we found that PGS\u003csub\u003eDYSdiscordant\u003c/sub\u003e was associated with worse performance in all subjects (OR\u0026thinsp;\u0026lt;\u0026thinsp;0.78, adj-P\u0026thinsp;\u0026le;\u0026thinsp;8.8E-15) and explained the largest proportion of variance on school performance of all genetic partitions, ranging from 1.76\u0026ndash;2.65% (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Conversely, PGS\u003csub\u003eDYSconcordant\u003c/sub\u003e was associated with better performance in mathematics (OR\u0026thinsp;=\u0026thinsp;1.10, 95% CI\u0026thinsp;=\u0026thinsp;1.04\u0026ndash;1.06), but dispayed no significant association in language-related subjects (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). PGS comparisons across ranked quintiles showed the expected trend of lower odds for better school performance in individuals with higher PGS\u003csub\u003eDYS\u003c/sub\u003e across all partitions but PGS\u003csub\u003eDYSconcordant\u003c/sub\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and Supplementary Table\u0026nbsp;6). This result is consistent with the probability plots for performance, where the PGS\u003csub\u003eDYSconcordant\u003c/sub\u003e partition tends to the same effect as the PGS\u003csub\u003eEA\u003c/sub\u003e (i.e., probability of better performance (grades A or B) increase with higher PGS\u003csub\u003eDYSconcordant\u003c/sub\u003e) while PGS\u003csub\u003eDYSdiscordant\u003c/sub\u003e was consistent with the results of PGS\u003csub\u003eDYS\u003c/sub\u003e (i.e., probability of worse performance (grades C or D) increased with higher PGS\u003csub\u003eDYSdiscordant\u003c/sub\u003e) (Supplementary Figure S4).\u003c/p\u003e\u003cp\u003eWhen stratifying by sex, no differences were found on the effect of the genetic liability of dyslexia dissected by its relationship with EA on school performance (Interaction P-value\u0026thinsp;\u0026gt;\u0026thinsp;0.05) (Supplementary Table\u0026nbsp;7).\u003c/p\u003e\u003cp\u003eTo assess whether the dyslexia diagnosis explained the association of the different PGS with school performance, we first confirmed in our in-house sample of 4,274 school-aged participants, including 469 diagnosed cases, that the dyslexia diagnosis was associated with PGS\u003csub\u003eDYSnoEA\u003c/sub\u003e (P\u0026thinsp;=\u0026thinsp;3.26E-04) and PGS\u003csub\u003eDYSdiscordant\u003c/sub\u003e (P\u0026thinsp;=\u0026thinsp;4.25E-09), but not with PGS\u003csub\u003eDYSconcordant\u003c/sub\u003e, which were subsequently considered for the mediation analysis (Supplementary Table\u0026nbsp;8A). We also found association between the dyslexia diagnosis and school performance in the three subjects under study (P\u0026le;1.07E-39) (Supplementary Table\u0026nbsp;8B), and confirmed that the effect of both PGS\u003csub\u003eDYSnoEA\u003c/sub\u003e and PGS\u003csub\u003eDYSdiscordant\u003c/sub\u003e on school performance were attenuated after considering the effect of the dyslexia diagnosis. These results suggest that dyslexia partially mediates the effect of these PGS on school performance (accounting for a reduction in effect size ranging from 10.44\u0026ndash;12.91%; Supplementary Tables\u0026nbsp;8C-D and Figure S5).\u003c/p\u003e\u003cp\u003e\u003cb\u003ePolygenic dissection of the contribution of dyslexia to childhood psychopathology by its relationship with EA\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe analysis of childhood psychopathology in our in-house INSchool cohort (n\u0026thinsp;=\u0026thinsp;4084 school-aged individuals with completed clinical information) revealed that PGS\u003csub\u003eEA\u003c/sub\u003e was negatively associated with attentional problems and externalizing behavior. In contrast, the genetic liability of dyslexia and all its SNPs subsets, but the concordant genomic partition, were positively associated with somatic complains as well as social and attentional problems. In addition, the SNP subset with discordant direction of the effect between dyslexia and EA was also associated with externalizing behavior, indicating that this genomic partition is associated with the majority of behavioral and emotional problems assessed (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and Supplementary Table\u0026nbsp;9).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eGenetic covariance between dyslexia and related disorders and traits\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe analysis of genetic covariance showed opposite directions of effect for concordant and discordant genomic partitions across most of the assessed phenotypes. The patterns observed for the DYS\u003csub\u003ediscordant\u003c/sub\u003e and DYS\u003csub\u003enoEA\u003c/sub\u003e partitions closely resembled the genetic correlations of dyslexia across the whole genome \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. In contrast, the concordant genomic partition, DYS\u003csub\u003econcordant\u003c/sub\u003e, exhibited a pattern more aligned with the genetic correlations observed for EA. We found that all traits related with cognition and language ability- except for ambidextry- as well as anorexia nervosa, autism spectrum disorder (ASD), substance use related traits (i.e., lifetime cannabis use and alcohol use), subjective well-being, and household income, displayed negative covariance with the discordant but positive covariance with the concordant genetic load for dyslexia. In contrast, several psychiatric disorders (i.e., ADHD, major depression, and anxiety disorder), along with the addiction factor, sleeplessness/insomnia and loneliness/isolation, showed positive covariance with the discordant and negative covariance with the concordant genetic variation. Interestingly, significant genetic covariance with dyslexia was observed for anxiety disorder and lifetime cannabis use only when the genetic liablity was dissected by genetic variants associated with EA (i.e., DYS\u003csub\u003ediscordant\u003c/sub\u003e and DYS\u003csub\u003econcordant\u003c/sub\u003e). Finally, we also found that the genetic covariance of dyslexia with some traits was not significant for some genomic partitions (i.e., ambidexterity, schizophrenia and bipolar disorder), and for general risk tolerance was independent of EA and in the same direction than those described for dyslexia across the whole genome (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e and Supplementary Table\u0026nbsp;10).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eThe results of this study confirmed negative genetic correlation between dyslexia and EA \u003csup\u003e\u003cspan additionalcitationids=\"CR8\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e and strong genetic overlap between them. Our findings also support the use of EA to dissect the genetic load of dyslexia, as previously described for schizophrenia \u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e, ADHD and ASD \u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e, to better understand the polygenic contribution of dyslexia to school performance and childhood psycopathology.\u003c/p\u003e\u003cp\u003eWe found that the genetic liability of EA and dyslexia were associated with school performance and showed opposite direction of the effect, as expected. PGS\u003csub\u003eEA\u003c/sub\u003e was strongly associated with better performance in primary language, foreign language and mathematics, while PGS\u003csub\u003eDYS\u003c/sub\u003e was associated with worse performance in all three subjects, as previously described in the literature \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. Accordingly, the association between the PGS\u003csub\u003eEA\u003c/sub\u003e and PGS\u003csub\u003eDYS\u003c/sub\u003e with childhood psychopathology showcased this opposite pattern of association, with dyslexia being associated with more somatic complaints and social and attentional problems, some of them previously described \u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eWhen dissecting the genetic liability of dyslexia by its relationship with EA, we found that all genomic partitions were associated with school performance in some of the studied academic subjects. In line with the negative genetic correlation found between dyslexia and EA, the discordant genetic partition, which accounts for the highest enrichment in dyslexia genetic background \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e, showed the strongest negative effect in school performance, which was partially mediated by the dyslexia diagnosis. In addition, this genomic partition was associated with higher rates of children psychopathology, with special mention to externalizing behavior. These results confirm that a high proportion of dyslexia risk loci is shared with EA and has opposite direction of the effect, and provide further evidence on the negative effects of dyslexia on long-term academic and emotional outcomes \u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. Conversely, PGS constructed with concordant variants were associated with better performance in mathematics. This genomic partition showed notably low SNP heritability for dyslexia given the proportion of SNPs included, was not associated with dyslexia diagnosis nor did it contribute to language skills or children's psychopathology in our sample, which supports that this partition might be more closely related to EA rather than to dyslexia itself. This is further supported by the results in genetic covariance, where the concordant genomic partition showed greater similarity to the pattern observed in EA.\u003c/p\u003e\u003cp\u003eNotably, most cognitive and psychiatric traits also exhibited opposite pattern of genetic covariance between the concordant and discordant genomic partitions of dyslexia, being the latter generaly associated with increased risk of mental disorders, reduced well-being, and lower cognitive performance, similarly to overall dyslexia variants described \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. These findings are consistent with recent studies showing a connection between poor academic achievement and a higher risk of subsequent mental disorders, with the exception of eating disorders, where the association is of similar magnitude but in the opposite direction \u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. We also found that both anxiety and lifetime cannabis use were associated with the polygenic load of dyslexia only when its genetic liability was dissected by its relationship with EA. These findings suggest that these traits are highly influenced by educational level and are consistent with previous studies exploring the relationship between dyslexia and anxiety \u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Overall, they reinforce the utility of EA-stratified analyses to uncover distinct genetic pathways within dyslexia and its comorbidities, and emphasize the importance of considering EA when studying the shared genetic underpinnings of cognitive and psychiatric outcomes.\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003eWe also found that the genetic liability of dyslexia not associated with EA has a negative impact on school performance, specially in first and foreing language. This suggests that the association between dyslexia and worse performance in language-related subjects is not only explained by shared genetic with EA, but also to specific dyslexia genetic liability independent of EA. Notably, this association was also partially mediated by dyslexia diagnosis and associated with higher rates of childhood psychopathology in our sample, following a pattern of association similar to the discordant genomic partition. These findings suggest that specific genetic components of dyslexia contribute to both language abilities and early psychopathologcal symptoms, beyond general cognitive or educational factors.\u003c/p\u003e\u003cp\u003eThis study aims to disentangle the polygenic contribution of dyslexia on school performance by incorporating a third trait, EA, facilitating a more comprehensive understanding of their polygenic effects and the early emergence of psychopathology in a well-characterized cohort. However, several considerations need to be taken into account. First, we observed that discordant polygenic signature between dyslexia and EA was associated with early-onset psychopathology and lower academic performance. This supports the existence of shared genetic factors that transcend diagnostic categories and suggests that early emerging symptoms may negatively impact school outcomes. However, we did not infer causality and cannot exclude reverse causation, where poor academic performance may itself increase the risk of developing psychopathology later in life \u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. This is specially relevant considering that there is evidence indicating that EA and school performance are influenced by demographic and indirect genetic effects \u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e, such as assortative mating and dynastic effects, which may lead to inaccurate estimations of direct genetic effects \u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e,\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. While the primary goal of this study was to explore the polygenic contributions of dyslexia to school performance in relation to EA, our findings should be cautiously interpreted and further family-based studies accounting for these population mechanisms are required. Secondly, despite accounting for confounding factors, residual confounders may persist. Diverse dyslexia treatments across schools, ranging from cognitive and behavioral therapies to classroom support and individualized educational plans \u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e, could influence school grades and potentially lead to an underestimation of polygenic effect of dyslexia on school performance. Lastly, analyses across the three academic subjects revelaed distinct patterns of association between language skills and mathematics. We found a stronger effect of the genetic load for dyslexia on language performance than on mathematics. Additionally, the genetic contribution on mathematics emerged only when considering the genomic patition shared with EA, whereas the dyslexia-specific genetic partition independent of EA contributed exclusively to school performance in language-related subjects. These findings support the existence of distinct genetic influences underlying language and mathematical abilities, highlighting the need for further studies to clarify these specific genetic contributions.\u003c/p\u003e\u003cp\u003eIn summary, we confirmed a negative genetic correlation between dyslexia and EA. By dissecting the genetic predisposition to dyslexia by its effect on EA, we found that the dyslexia genetic liability independent of EA was associated with poorer performance in language-related subjects and increased childhood psychopathology, supporting the existence of dyslexia-specific genetic effects beyond general cognitive or educational pathways. Moreover, the dyslexia genetic components associated with EA exhibited opposing effects: the concordant partition was associated with better performance only in mathematics, while the discordant partition was associated with worse school grades in all the studied subjects, partially mediated by dyslexia diagnosis, higher rates of children psychopathology, reduced well-being, and lower cognitive performance - mirroring the overall dyslexia signal. These findings reinforce the utility of genetic liability for EA as a valuable tool for understanding the genetic interplay between dyslexia and academic outcomes.\u003c/p\u003e\u003cp\u003e\u003cb\u003eETHICAL INFORMATION\u003c/b\u003e\u003c/p\u003e\u003cp\u003e The project was approved by the Ethics Committee at the Hospital Universitari Vall d\u0026rsquo;Hebron (PR(AG)491/2022) and the CEIm Fundaci\u0026oacute; Sant Joan de D\u0026eacute;u (PIC-154-22) and written informed consent was obtained from parents or caregivers.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003ch2\u003eDECLARATION OF INTERESTS\u003c/h2\u003e\u003cp\u003eJ.A.R.Q was on the speakers\u0026rsquo; bureau and/or acted as consultant for Biogen, Janssen-Cilag, Novartis, Shire, Takeda, Bial, Shionogi, Sincrolab, Novartis, BMS, Medice, Rubi\u0026oacute;, Uriach, Technofarma and Raffo in the last 3 years. He also received travel awards (air tickets\u0026thinsp;+\u0026thinsp;hotel) for taking part in psychiatric meetings from Janssen-Cilag, Rubi\u0026oacute;, Shire, Takeda, Shionogi, Bial and Medice. The Department of Psychiatry chaired by him received unrestricted educational and research support from the following companies in the last 3 years: Janssen- Cilag, Shire, Oryzon, Roche, Psious, and Rubi\u0026oacute;. The rest of authors have nothing to disclose.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFUNDING\u003c/h2\u003e\u003cp\u003eThis work was supported by the Ag\u0026egrave;ncia de Gesti\u0026oacute; d\u0026rsquo;Ajuts Universitaris i de Recerca (AGAUR, 2017SGR-1461, 2021SGR-00840, 2021SGR-01093); the Instituto de Salud Carlos III (PI20/00041, PI22/00464, PI23/00404, PI23/00026, PI24/00195, CP22/00128 to M.S.A, CP22/00026 to S.A, FI23/00152 to P.C.G); the Network Center for Biomedical Research (CIBER) to J.C.D and U.Z.A.; the European Regional Development Fund (ERDF); the ECNP Network \u0026lsquo;ADHD across the Lifespan\u0026rsquo;; \u0026ldquo;la Marat\u0026oacute; de TV3\u0026rdquo; (202228-30 and 202228-31); the European Union H2020 Programme (H2020/2014\u0026ndash;2020) under grant agreements no. 848228 (DISCOvERIE) and no. 2020604 (TIMESPAN); \u0026ldquo;Fundaci\u0026oacute; \u0026lsquo;la Caixa\u0026rsquo;, Diputaci\u0026oacute; de Barcelona, Pla Estrat\u0026egrave;gic de Recerca i Innovaci\u0026oacute; en Salut\u0026rdquo; (PERISSLT006/17/285); \u0026ldquo;Fundaci\u0026oacute; Privada d'Investigaci\u0026oacute; Sant Pau\u0026rdquo; (FISP); Ministry of Health of Generalitat de Catalunya; grant RYC2021-033573-I funded by MICIU/AEI/\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.13039/501100011033\u003c/span\u003e\u003cspan address=\"10.13039/501100011033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e and European Union NextGenerationEU/PRTR to MM.\u003c/p\u003e\u003ch2\u003eACKNOWLEDGMENTS\u003c/h2\u003e\u003cp\u003eThe authors are grateful to families, students, and staff of the public primary schools (i.e., Joan Maragall, Mar\u0026iacute;a Bores, Marqu\u0026eacute;s de la Pobla, Martinet, Pins del Vall\u0026egrave;s, Puiggracios, Sant Jordi, Ramon Llull, Rivo Rubeo, Tagamanent and Teresa Bergued\u0026agrave;), public secondary schools (i.e., Angeleta Ferrer i Sensat, Antoni Pous i Argila, Cal Gravat, Duc de Montblanc, Institut del Ter, Jaume Call\u0026iacute;s, Lacet\u0026agrave;nia, Llu\u0026iacute;s de Peguera, Mol\u0026iacute; de la Vila, Montsuar, Pius Font i Quer, Vallbona d\u0026rsquo;Anoia, and Vil\u0026ensp;la Romana), and private schools (i.e., Airina, L'Ave Maria, Casals \u0026ndash; Gr\u0026agrave;cia, Episcopal Lleida, La Farga, FEDAC Manresa, FEDAC Vic, Garb\u0026iacute; Pere Verg\u0026eacute;s Esplugues, Institucio Igualada, Joviat, Oms i de Prat, Pies Matar\u0026oacute;, Pureza de Maria, Regina Carmeli, Sagrats Cors Centelles, La Salle Manlleu, La Salle Manresa, Sant Miquel dels Sants, Thau Barcelona and Vedruna Escorial Vic) who kindly contribute in this research. We would like to thank the research participants and employees of 23andMe, Inc. for making this work possible. The genotyping service was carried out at the Genotyping Unit-CEGEN in the Spanish National Cancer Research Centre (CNIO), supported by Instituto de Salud Carlos III (ISCIII), Ministerio de Ciencia e Innovaci\u0026oacute;n. CEGEN is part of the initiative IMPaCTGEN\u0026oacute;MICA (IMP/00009) cofunded by ISCIII and the European Regional Development Fund (ERDF).\u003c/p\u003e\u003ch2\u003eDATA AVAILABILITY\u003c/h2\u003e\u003cp\u003eData from the INSchool cohort included in this article is not publicly available due to limitations in ethical approvals and the summary data will be available upon reasonable request. The full GWAS summary statistics from the original 23andMe discovery studies have been made available through 23andMe to qualified researchers under agreements with 23andMe that protects participants privacy. Datasets will be made available at no cost for academic use, visit \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://research.23andme.com/collaborate/#dataset-access/\u003c/span\u003e\u003cspan address=\"https://research.23andme.com/collaborate/#dataset-access/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e for more information and to apply to access. All participants from 23andMe provided informed consent and volunteered to participate in the research online, under a protocol approved by the external AAHRPP-accredited insttutional review board, Ethical and Independent Review Services. As of 2022, Ethical and Independent Review Services is part of Salus institutional review board (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.versiticlinicaltrials.org/salusirb\u003c/span\u003e\u003cspan address=\"https://www.versiticlinicaltrials.org/salusirb\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eErbeli F, Rice M, Paracchini S. Insights into Dyslexia Genetics Research from the Last Two Decades. \u003cem\u003eBrain Sci\u003c/em\u003e 2021; 12: 27.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWagner RK, Zirps FA, Edwards AA, Wood SG, Joyner RE, Becker BJ \u003cem\u003eet al.\u003c/em\u003e The Prevalence of Dyslexia: A New Approach to Its Estimation. \u003cem\u003eJ Learn Disabil\u003c/em\u003e 2020; 53: 354\u0026ndash;365.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWilmot A, Pizzey H, Leit\u0026atilde;o S, Hasking P, Boyes M. Growing up with dyslexia: Child and parent perspectives on school struggles, self-esteem, and mental health. \u003cem\u003eDyslexia\u003c/em\u003e 2023; 29: 40\u0026ndash;54.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCiulkinyte A, Mountford HS, Fontanillas P, Team R, Bates TC, Martin NG \u003cem\u003eet al.\u003c/em\u003e Genetic neurodevelopmental clustering and dyslexia. \u003cem\u003emedRxiv\u003c/em\u003e 2023. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1101/2023.10.04.23296530\u003c/span\u003e\u003cspan address=\"10.1101/2023.10.04.23296530\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHaworth CMA, Kovas Y, Harlaar N, Hayiou-Thomas ME, Petrill SA, Dale PS \u003cem\u003eet al.\u003c/em\u003e Generalist genes and learning disabilities: A multivariate genetic analysis of low performance in reading, mathematics, language and general cognitive ability in a sample of 8000 12-year-old twins. \u003cem\u003eJ Child Psychol Psychiatry\u003c/em\u003e 2009; 50: 1318\u0026ndash;1325.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDoust C, Fontanillas P, Eising E, Gordon SD, Wang Z, Alag\u0026ouml;z G \u003cem\u003eet al.\u003c/em\u003e Discovery of 42 genome-wide significant loci associated with dyslexia. \u003cem\u003eNat Genet\u003c/em\u003e 2022; 54: 1621\u0026ndash;1629.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVerhoef E, Demontis D, Burgess S, Shapland CY, Dale PS, Okbay A \u003cem\u003eet al.\u003c/em\u003e Disentangling polygenic associations between attention-deficit/hyperactivity disorder, educational attainment, literacy and language. \u003cem\u003eTransl Psychiatry\u003c/em\u003e 2019; 9. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41398-018-0324-2\u003c/span\u003e\u003cspan address=\"10.1038/s41398-018-0324-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLuciano M, Hagenaars SP, Cox SR, Hill WD, Davies G, Harris SE \u003cem\u003eet al.\u003c/em\u003e Single Nucleotide Polymorphisms Associated with Reading Ability Show Connection to Socio-Economic Outcomes. \u003cem\u003eBehav Genet\u003c/em\u003e 2017; 47: 469\u0026ndash;479.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRichardson JTE, Wydell TN. The representation and attainment of students with dyslexia in UK higher education. 2003.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee JJ, Wedow R, Okbay A, Kong E, Maghzian O, Zacher M \u003cem\u003eet al.\u003c/em\u003e Gene discovery and polygenic prediction from a genome-wide association study of educational attainment in 1.1 million individuals. \u003cem\u003eNat Genet\u003c/em\u003e 2018; 50: 1112\u0026ndash;1121.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBranigan AR, Mccallum KJ, Freese J. Variation in the heritability of educational attainment: An international meta-analysis. \u003cem\u003eSocial Forces\u003c/em\u003e 2013; 92: 109\u0026ndash;140.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCutler D, Lleras-Muney A. Education and Health: Evaluating Theories and Evidence. Cambridge, MA, 2006 doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3386/w12352\u003c/span\u003e\u003cspan address=\"10.3386/w12352\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKrapohl E, Rimfeld K, Shakeshaft NG, Trzaskowski M, McMillan A, Pingault JB \u003cem\u003eet al.\u003c/em\u003e The high heritability of educational achievement reflects many genetically influenced traits, not just intelligence. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e 2014; 111: 15273\u0026ndash;15278.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWard ME, McMahon G, St Pourcain B, Evans DM, Rietveld CA, Benjamin DJ \u003cem\u003eet al.\u003c/em\u003e Genetic variation associated with differential educational attainment in adults has anticipated associations with school performance in children. \u003cem\u003ePLoS One\u003c/em\u003e 2014; 9. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0100248\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0100248\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLeit\u0026atilde;o S, Dzidic P, Claessen M, Gordon J, Howard K, Nayton M \u003cem\u003eet al.\u003c/em\u003e Exploring the impact of living with dyslexia: The perspectives of children and their parents. \u003cem\u003eInt J Speech Lang Pathol\u003c/em\u003e 2017; 19: 322\u0026ndash;334.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSinger E. Coping with academic failure, a study of Dutch children with dyslexia. \u003cem\u003eDyslexia\u003c/em\u003e 2008; 14: 314\u0026ndash;333.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCarawan LW, Nalavany BA, Jenkins C. Emotional experience with dyslexia and self-esteem: The protective role of perceived family support in late adulthood. \u003cem\u003eAging Ment Health\u003c/em\u003e 2016; 20: 284\u0026ndash;294.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHumphrey N, Mullins PM. Self-concept and self‐esteem in developmental dyslexia. \u003cem\u003eJournal of Research in Special Educational Needs\u003c/em\u003e 2002; 2. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1471-3802.2002.00163.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1471-3802.2002.00163.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBosch R, Pagerols M, Rivas C, Sixto L, Bricoll\u0026eacute; L, Espa\u0026ntilde;ol-Mart\u0026iacute;n G \u003cem\u003eet al.\u003c/em\u003e Neurodevelopmental disorders among Spanish school-age children: prevalence and sociodemographic correlates. \u003cem\u003ePsychol Med\u003c/em\u003e 2022; 52: 3062\u0026ndash;3072.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCuetos F, Rodr\u0026iacute;guez B, Ruano E, Arribas D. \u003cem\u003ePROLEC-R: Bater\u0026iacute;a de evaluaci\u0026oacute;n de los procesos lectores. Revisada.\u003c/em\u003e. Madrid: TEA Ediciones., 2007.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCuetos F, Arribas D, Ramos JL. \u003cem\u003ePROLEC-SE-R. Bater\u0026iacute;a para la Evaluaci\u0026oacute;n de los Procesos Lectores en Secundaria y Bachillerato - Revisada.\u003c/em\u003e Madrid: TEA Ediciones, 2016\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://web.teaediciones.com/PROLEC-SE-R-Bateria-para-la-Evaluacion-de-los-Procesos-Lectores-en-Secundaria-y-Bachillerato\u0026mdash;Revisada.aspx\u003c/span\u003e\u003cspan address=\"https://web.teaediciones.com/PROLEC-SE-R-Bateria-para-la-Evaluacion-de-los-Procesos-Lectores-en-Secundaria-y-Bachillerato\u0026mdash;Revisada.aspx\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (accessed 2 Sep2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWechsler D. \u003cem\u003eWechsler intelligence scale for children\u0026ndash;fifth edition (WISC-V).\u003c/em\u003e. San Antonio, TX: NCS Pearson, 2014.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBosch R, Pagerols M, Rivas C, Sixto L, Bricoll\u0026eacute; L, Espa\u0026ntilde;ol-Mart\u0026iacute;n G \u003cem\u003eet al.\u003c/em\u003e Neurodevelopmental disorders among Spanish school-age children: prevalence and sociodemographic correlates. \u003cem\u003ePsychol Med\u003c/em\u003e 2022; 52: 3062\u0026ndash;3072.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAchenback TM, Rescorla LA. \u003cem\u003eManual for ASEBA school-age forms \u0026amp; profiles\u003c/em\u003e. University of Vermont, Research Center for Children, Youth \u0026amp; Families, 2001.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCabana-Dom\u0026iacute;nguez J, Bosch R, Soler Artigas M, Alemany S, Llonga N, Vilar-Rib\u0026oacute; L \u003cem\u003eet al.\u003c/em\u003e Dissecting the polygenic contribution of attention-deficit/hyperactivity disorder and autism spectrum disorder on school performance by their relationship with educational attainment. \u003cem\u003eMolecular Psychiatry 2024\u003c/em\u003e 2024;: 1\u0026ndash;13.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDas S, Forer L, Sch\u0026ouml;nherr S, Sidore C, Locke AE, Kwong A \u003cem\u003eet al.\u003c/em\u003e Next-generation genotype imputation service and methods. \u003cem\u003eNat Genet\u003c/em\u003e 2016; 48: 1284\u0026ndash;1287.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChoi SW, Mak TS-H, O\u0026rsquo;Reilly PF. Tutorial: a guide to performing polygenic risk score analyses. \u003cem\u003eNat Protoc\u003c/em\u003e 2020; 15: 2759\u0026ndash;2772.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAndreu-Bernabeu \u0026Aacute;, D\u0026iacute;az-Caneja CM, Costas J, De Hoyos L, Stella C, Gurriar\u0026aacute;n X \u003cem\u003eet al.\u003c/em\u003e Polygenic contribution to the relationship of loneliness and social isolation with schizophrenia. \u003cem\u003eNat Commun\u003c/em\u003e 2022; 13. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/S41467-021-27598-6\u003c/span\u003e\u003cspan address=\"10.1038/S41467-021-27598-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBulik-Sullivan B, Loh PR, Finucane HK, Ripke S, Yang J, Patterson N \u003cem\u003eet al.\u003c/em\u003e LD score regression distinguishes confounding from polygenicity in genome-wide association studies. \u003cem\u003eNat Genet\u003c/em\u003e 2015; 47: 291\u0026ndash;295.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFrei O, Holland D, Smeland OB, Shadrin AA, Fan CC, Maeland S \u003cem\u003eet al.\u003c/em\u003e Bivariate causal mixture model quantifies polygenic overlap between complex traits beyond genetic correlation. \u003cem\u003eNat Commun\u003c/em\u003e 2019; 10. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-019-10310-0\u003c/span\u003e\u003cspan address=\"10.1038/s41467-019-10310-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFinucane HK, Bulik-Sullivan B, Gusev A, Trynka G, Reshef Y, Loh PR \u003cem\u003eet al.\u003c/em\u003e Partitioning heritability by functional annotation using genome-wide association summary statistics. \u003cem\u003eNat Genet\u003c/em\u003e 2015; 47: 1228\u0026ndash;1235.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChristensen RHB. Regression Models for Ordinal Data [R package ordinal version 2023.12-4]. 2023.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFox J. Effect Displays in R for Generalised Linear Models. \u003cem\u003eJ Stat Softw\u003c/em\u003e 2003; 8. doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.18637/jss.v008.i15\u003c/span\u003e\u003cspan address=\"10.18637/jss.v008.i15\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHollingshead AB. \u003cem\u003eFour Factor Index of Social Status\u003c/em\u003e. Yale Journal of Sociology, 2011.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLu Q, Li B, Ou D, Erlendsdottir M, Powles RL, Jiang T \u003cem\u003eet al.\u003c/em\u003e A Powerful Approach to Estimating Annotation-Stratified Genetic Covariance via GWAS Summary Statistics. \u003cem\u003eAm J Hum Genet\u003c/em\u003e 2017; 101: 939\u0026ndash;964.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBansal V, Mitjans M, Burik CAP, Linn\u0026eacute;r RK, Okbay A, Rietveld CA \u003cem\u003eet al.\u003c/em\u003e Genome-wide association study results for educational attainment aid in identifying genetic heterogeneity of schizophrenia. \u003cem\u003eNature Communications 2018 9:1\u003c/em\u003e 2018; 9: 1\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGialluisi A, Andlauer TFM, Mirza-Schreiber N, Moll K, Becker J, Hoffmann P \u003cem\u003eet al.\u003c/em\u003e Genome-wide association study reveals new insights into the heritability and genetic correlates of developmental dyslexia. \u003cem\u003eMol Psychiatry\u003c/em\u003e 2021; 26: 3004\u0026ndash;3017.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDahle AE, Knivsberg AM. Internalizing, externalizing and attention problems in dyslexia. \u003cem\u003eScandinavian Journal of Disability Research\u003c/em\u003e 2014; 16: 179\u0026ndash;193.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLivingston EM, Siegel LS, Ribary U. Developmental dyslexia: emotional impact and consequences. \u003cem\u003eAust J Learn Diffic\u003c/em\u003e 2018; 23: 107\u0026ndash;135.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWeckstr\u0026ouml;m T, Elovainio M, Pulkki-R\u0026aring;back L, Suokas K, Komulainen K, Mullola S \u003cem\u003eet al.\u003c/em\u003e School achievement in adolescence and the risk of mental disorders in early adulthood: a Finnish nationwide register study. \u003cem\u003eMol Psychiatry\u003c/em\u003e 2023; 28: 3104\u0026ndash;3110.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNovita S. Secondary symptoms of dyslexia: a comparison of self-esteem and anxiety profiles of children with and without dyslexia. \u003cem\u003eEur J Spec Needs Educ\u003c/em\u003e 2016; 31: 279\u0026ndash;288.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBrittain H, Vaillancourt T. Longitudinal associations between academic achievement and depressive symptoms in adolescence: Methodological considerations and analytical approaches for identifying temporal priority. In: \u003cem\u003eAdvances in Child Development and Behavior\u003c/em\u003e. Academic Press Inc., 2023, pp 327\u0026ndash;355.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHowe LJ, Nivard MG, Morris TT, Hansen AF, Rasheed H, Cho Y \u003cem\u003eet al.\u003c/em\u003e Within-sibship genome-wide association analyses decrease bias in estimates of direct genetic effects. \u003cem\u003eNat Genet\u003c/em\u003e 2022; 54: 581\u0026ndash;592.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNivard MG, Belsky DW, Harden KP, Baier T, Andreassen OA, Ystr\u0026oslash;m E \u003cem\u003eet al.\u003c/em\u003e More than nature and nurture, indirect genetic effects on children\u0026rsquo;s academic achievement are consequences of dynastic social processes. \u003cem\u003eNat Hum Behav\u003c/em\u003e 2024; 8: 771\u0026ndash;778.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHugh-Jones D, Verweij KJH, St. Pourcain B, Abdellaoui A. Assortative mating on educational attainment leads to genetic spousal resemblance for polygenic scores. \u003cem\u003eIntelligence\u003c/em\u003e 2016; 59: 103\u0026ndash;108.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGeredakis A, Vergou M, Zakopoulou V. Complementary and Alternative Approaches to Therapeutic Dyslexia Intervention. \u003cem\u003eWorld J Res Rev\u003c/em\u003e 2017; 4: 36\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDuff DM, Hendricks AE, Fitton L, Adlof SM. Reading and Math Achievement in Children With Dyslexia, Developmental Language Disorder, or Typical Development: Achievement Gaps Persist From Second Through Fourth Grades. \u003cem\u003eJ Learn Disabil\u003c/em\u003e 2023; 56: 371\u0026ndash;391.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAshkenazi S, Black JM, Abrams DA, Hoeft F, Menon V. Neurobiological Underpinnings of Math and Reading Learning Disabilities. \u003cem\u003eJ Learn Disabil\u003c/em\u003e 2013; 46: 549\u0026ndash;569.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"translational-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"tp","sideBox":"Learn more about [Translational Psychiatry](http://www.nature.com/tp/)","snPcode":"41398","submissionUrl":"https://mts-tp.nature.com/cgi-bin/main.plex","title":"Translational Psychiatry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"learning disorder, educational attainment, polygenic risk scores, shared genetics","lastPublishedDoi":"10.21203/rs.3.rs-7206327/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7206327/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDyslexia is a neurodevelopmental disorder typically diagnosed in school-aged children and associated with poor school performance and lower levels of educational attainment (EA). Here, we aimed to test the shared genetic architecture between dyslexia and EA, to dissect the polygenic contribution of dyslexia by its relationship with EA and to assess how these genetic partitions influence school performance, early manifestation of psychopathology and related traits. We first confirmed a negative genetic correlation between dyslexia and EA (rg=-0.186, SE\u0026thinsp;=\u0026thinsp;0.019, P\u0026thinsp;=\u0026thinsp;1.75E-22). Then, polygenic scores for EA and dyslexia were tested in a cohort of 4,274 school-aged children, revealing opposite direction of the effect in school performance. Next, we dissected the genetic liability for dyslexia into components shared with, and independent of, EA. The results revealed similar patterns of association for performance in primary and foreign languages, but distinct patterns when comparing these language-related subjects with mathematics. The dyslexia-specific genetic component independent of EA was associated with poorer academic outcomes in language-related subjects and increased rates of psychopathology, supporting the existence of dyslexia-specific genetic effects beyond general cognitive or educational pathways. In contrast, the genetic load of dyslexia that overlaps with EA contribute to school performance in both language-related subjects and mathematics and displayed opposite patterns of association dependent on whether concordant and discordant genomic partitions were considered. The discordant partition was associated with poorer school performance and higher rates of behavioral and emotional problems, being these associations partially mediated by the dyslexia diagnosis (accounting for a reduction in effect size ranging from 10.44 to 12.91%). Conversely, the concordant partition was only associated with better performance in mathematics. Overall, these findings highlight the polygenic contribution of dyslexia to both academic and psychopathological outcomes, support distinct genetic influences on language skills and mathematics, and uncerscore the usage of the genetic load for EA to deepen insight into the complex genetic relationship between dyslexia and school performance.\u003c/p\u003e","manuscriptTitle":"Towards Disentangling the Polygenic Contribution of Dyslexia to School Performance","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-20 06:42:42","doi":"10.21203/rs.3.rs-7206327/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"revise","date":"2026-01-13T13:07:40+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2025-12-22T12:26:26+00:00","index":2,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-12-10T09:54:18+00:00","index":2,"fulltext":"This content is not available."},{"type":"editorInvitedReview","content":"This content is not available.","date":"2025-08-13T10:11:06+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewerAgreed","content":"This content is not available.","date":"2025-08-12T10:04:34+00:00","index":1,"fulltext":"This content is not available."},{"type":"reviewersInvited","content":"","date":"2025-08-11T18:31:58+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-28T10:02:53+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-07-28T09:48:46+00:00","index":"","fulltext":""},{"type":"submitted","content":"Translational Psychiatry","date":"2025-07-25T11:35:08+00:00","index":"","fulltext":""},{"type":"checksFailed","content":"","date":"2025-07-25T10:59:39+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"translational-psychiatry","isNatureJournal":false,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"tp","sideBox":"Learn more about [Translational Psychiatry](http://www.nature.com/tp/)","snPcode":"41398","submissionUrl":"https://mts-tp.nature.com/cgi-bin/main.plex","title":"Translational Psychiatry","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"228bc5c8-f7f5-44e2-bc14-8d123ec987f3","owner":[],"postedDate":"August 20th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":53403275,"name":"Biological sciences/Genetics/Genomics"},{"id":53403276,"name":"Health sciences/Diseases/Psychiatric disorders"}],"tags":[],"updatedAt":"2026-05-13T09:59:55+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-20 06:42:42","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7206327","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7206327","identity":"rs-7206327","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}