Cumulative genetic score of KIAA0319 affects reading fluency through parental education levels and rapid automatized naming

doi:10.21203/rs.3.rs-2099080/v1

Download PDF

Research Article

Cumulative genetic score of KIAA0319 affects reading fluency through parental education levels and rapid automatized naming

https://doi.org/10.21203/rs.3.rs-2099080/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 01 Jun, 2023

Read the published version in Behavioral and Brain Functions →

You are reading this latest preprint version

KIAA0319, a well-studied candidate gene, has been shownto be associated with dyslexia and reading ability. In the present study, we investigated whether KIAA0319 interacts with the parental education level and whether rapid automatized naming (RAN), phonological awareness and morphological awareness mediate the relationship between the gene and reading ability. A total of 13 SNPs were selected, and the cumulative genetic risk score (CGS) was calculated. The interaction effect suggested that individuals with a low CGS were better at fluent reading in a positive environment than individuals with a high CGS, but not accurate reading. Moreover, the interaction coincided with the differential susceptibility model. The results of the multiple mediator model revealed that digit RAN mediates the impact of the genetic cumulative effect on fluent reading. These findings provide evidence that KIAA0319 is a risk vulnerability gene that interacts with environmental factorsto impact reading ability and demonstratethe reliability of RAN as an endophenotype.

KIAA0319

reading ability

mediation

endophenotypes

vantage sensitivity model

differential susceptibility model

Reading ability is a complex behavioral characteristic that mainly includes reading accuracy and reading fluency. The acquisition of these skills involves a series of cognitive processes, such as phonological awareness, orthographic rules, morpheme awareness and rapid automatized naming [1]. Currently, the variation in reading ability is largely due to genetic factors: genetic variation accounts for 20–80% of the total variation in reading skills, and the genetic variations discovered thus far only explain the “tip of the iceberg” of estimated heritability [1, 2].

Among the susceptibility genes identified to date, KIAA0319 is an important candidate gene. Schmitz et al. [3]analyzed DNA methylation in the KIAA0319 (a gene involved in dyslexia and ciliogenesis) promoter region to investigate whether epigenetic markers of language lateralization could be identified in nonneuronal tissue. KIAA0319 affects neuronal migration, neurite outgrowth, cortical morphogenesis, and ciliary structure and function [4–7]. These data provide a framework to interpret the effects of the dyslexia-associated genetic variants that reside in KIAA0319 noncoding regulatory regions [8].

Franks et al. [9] used quantitative trait loci (QTL) analysis and found that single nucleotide polymorphisms (SNPs) were associated with reading-related traits, such as word reading, phonological awareness, and orthographic rules. Moreover, rs2143340 was replicated in the Avon Longitudinal Study of Parents and Children (ALSPAC) and found to be associated with poor performance on continuous measures of reading and spelling [10]. In subsequent studies, Scerri et al. [11] found that rs2143340 in KIAA0319 was significantly correlated with word reading accuracy in their quantitative analysis of reading phenotypes. Quantitative analysis of samples from Dutch children with developmental dyslexia found that rs761100 and rs2038137 in KIAA0319 were related to digit rapid automatized naming, and that rs6935076 was related to word reading fluency [12]. However, other studies have not found a relationship between genes and reading ability [13, 14].

These inconsistencies can be explained in part by heterogeneity, possibly due to the different criteria for phenotypic assessment, age, sample size, population genetic background, and neglect of the environment. However, when studying the role of candidate genes, endophenotypes (EPs) can reduce the effect of heterogeneity, and environmental factors can increase the probability of gene expression in behavior [15].

Endophenotypes are more useful than macroscopic behavior indicators and serve as mediating variables for understanding complex pathways [16, 17]. EPs reflect neurophysiological, biochemical, endocrinological, cognitive or neuropsychological processes that are associated with traits or diseases and may reflect specific genes that are associated with behavioral phenotypes [18–20].

Cognitive skill is an important factor in the development of children's reading abilities. Rapid automatized naming (RAN), a well-studied cognitive process, has been widely used in studies of reading acquisition and has been found to be reliably related to reading achievement [21, 22]. Several studies have identified RAN to be independent of reading and reading-related cognitive processes, such as orthographic processing, phonological awareness, and short-term memory [23–25]. Moreover, RAN has been found to be heritable. Genome-wide association studies reported the effects of SNPs on RAN, and classical twin studies found heritability estimates ranging from 0.56 to 0.70 for RAN [26–30]. The latest results explored RNA as an endophenotype that mediates the association between genes and reading ability [14, 31]. Therefore, RNA, according to the criteria for EPs [16, 32], was investigated as a mediator in the present study.

Phonological awareness and morphological awareness contribute to the development of reading ability and significantly predict word recognition and reading speed [33–36]. However, as cognitive skills related to reading, no studies have explored the possibility of phonological awareness and morphological awareness as endophenotypes. The heritability estimates of phonological awareness and morphological awareness were 0.19–0.83 [28, 37, 38] and 0.44–0.55 [39], respectively, in twin studies. Candidate gene studies have also discussed the association of KIAA0319 with phonological and morphological awareness [40, 41]. Therefore, this study is the first to investigate the mediating effect of phonological awareness and morphological awareness as intermediate phenotypes.

Studies of gene‒environment interactions have explored several environmental factors, e.g., the home literacy environment (HLE), socioeconomic status (SES), the prenatal environment, and computer game interventions [14, 42–44]. For KIAA0319, the presence of an interaction between maternal stress during pregnancy and the rs12193738 polymorphism was shown to affect reading ability at 16 years of age [42]. In gene×SES interactions, DYX1C1 interacted with SES on reading ability. SES is a well-explored variable in reading and dyslexia [45, 46]. Therefore, parental education was viewed as a proxy for SES in our study.

Three gene‒environment models exist at present. The diathesis-stress model emphasizes that individuals with disease risk alleles or vulnerability genotypes have higher negative environmental sensitivity [47, 48]. The vantage sensitivity model aims to describe the different responses of individuals in positive environments [48]. According to the differential susceptibility model, the same genetic traits or genotypes also have the effect of making individuals better (for better) or worse (for worse) under environmental factors [49]. In the current study, the intersection point of the interaction was used to determine which model the G×E term was more consistent with. To explore whether KIAA0319 should be considered a vulnerable or plastic gene [50], we conducted interaction and competing model analyses.

As a preliminary finding, D'Souza et al. [42] discussed the two polymorphisms in KIAA0319 affecting individual sensitivity to maternal stress. However, one polymorphism cannot represent the variation in gene function, and many variants with small effects cannot be detected. The cumulative genetic score (CGS) can be used to investigate the influence of multiple genetic polymorphisms [51–54]. The CGS was combined into a single score by assigning points (0, 1 or 2) according to the number of sensitive alleles [55]. Using the CGS can increase statistical power and model more variants on a gene. In the present study, we used the CGS approach to explore whether the CGS of KIAA0319 interacts with the parental education level and the mediating role of RNA between the CGS and fluent reading.

In this study, we selected the previously reported SNP of KIAA0319 in the Chinese population. We wanted to investigate (1) the effect of the KIAA0319 interaction with the environment on reading abilities; (2) whether genes affect reading abilities through endophenotypes; and (3) which gene‒environmental interaction model better fits the GXE effect.

Participants

A total of 2284 subjects (primary school students from Shaanxi, Gansu and Inner Mongolia) were recruited. All subjects were normal school-age students from grade 3 to grade 6 without any history of mental illness. In this study, saliva samples were collected (gene samples were extracted). In addition, some of the participants completed six endophenotype tests and two behavioral phenotype tests, which included Chinese character recognition, fluent reading and reading-related cognitive skills.

Genotyping And Snp Selection

An Illumina Asian Screening Array (ASA, 700K-750K) chip was used for genotyping the obtained DNA samples using the Beijing Compass biotechnology formula. PLINK was used to screen the following genotypes for standard quality control (Aderson et al., 2010; Chang, 2015): a single sample SNP detection rate higher than 0.9 (sample call rate > .90); single SNP detection rates (SNP call rate > .95); Hardy-Weinberg equilibrium coefficients (p < 10^− 5); minor allele frequencies (MAF > .01); and those with first degree of kinship was removed (PI_HAT > .50). MACH 4.0 software was used to carry out full genome data (imputation) analysis based on the Asian population data in the Genome Asia Pilot (GAsp) project, and the filled data were consistent with the previous quality control standards. Thirteen SNPs were extracted by PLINK v1.90.

Phenotypes

Word list reading task. As a measure of fluent reading, 1588 subjects read a word list consisting of 180 words, all of which were simple and common words. The subjects were asked to read the words quickly in order (both words of two words were correctly marked as 1, otherwise they were marked as 0). Finally, the time the subjects spent reading was used as a test indicator of fluent reading.

Rapid automatized naming task. The rapid automatized naming task includes four tasks: digit naming, picture naming, dice naming, and color naming. Each task table contains eight rows with five items in each row. Each naming task requires the participants to name the items in each row as fast as possible: five figures (2, 4, 6, 7, and 9), five dice pictures (5, 4, 2, 3 and 1), five pictures (dog, flower, book, shoe, and window), and five color pictures (red, yellow, black, green, and blue). Finally, the average time of two rapid automatized naming tasks is taken as the standard measurement. The test-retest reliabilities of the rapid automatized digit, picture, dice, and color naming tasks are 0.87, 0.82, 0.74, and 0.74, respectively.

Phonological awareness task. In this task, a child is verbally presented with a one-syllable word. The child’s task is to remove a given phoneme from the syllable in the word. The task consists of 16 items: initial phoneme deletion items (e.g., /mei4/ (sister) without /m/), middle phoneme deletion items (e.g., /tuan4/ (group) without /u/), and final phoneme deletion items (e.g., /guan1/ (close) without /n/). Scores are calculated if all tasks are pitched and pronounced correctly. This task has been widely used in language studies of Chinese children [35, 56, 57].

Morphological awareness task. Children are asked to identify one of the morphemes among two-morpheme words and to create two new words with the target morpheme. One of the morphemes in two words has the same meaning as the target morpheme; conversely, one of the morphemes in the other two words has a different meaning. Presented with the word bei1bao1 (which means backpack), children are asked to produce two new words containing [bao1], which has the same meaning as "bei1bao1", such as shu1bao1 (which means bag). At the same time, the meaning of [bao1] in two new words is different from that of bei1bao1, such as bao1zi1 (which means steamed stuffed bun).

Parental Education (Pe) Levels

A total of 1620 subjects in this study had information about their parents' educational levels, with 1 representing the lowest educational level and 8 representing the highest educational level. When both parents had access to information, the average score was determined by their educational level, which represented the education level of the parents. The average score was given according to the structure of the Chinese education system: 1 = primary school education, 2 = junior high school education, 3 = senior high school education, 4 = junior college education, 5 = undergraduate degree, 6 = master’s degree, 7 = doctoral degree, and 8 = postdoc.

Data analysis

SNP coding and cumulative genetic scores (CGSs)

The SNPs on KIAA0319 were repeated in GWAS in the Chinese sample [58]. We therefore adopted the β value for the phenotype of fluent reading and character recognition in this study. Coding was based on the first allele and beta value. When the value of β was positive, the homozygous with the first allele was 2 and the heterozygous was 1, 0 was the homozygous for the minor allele. When the beta value was negative, it was opposite to the encoding genotype. Finally, the genetic risk score of the KIAA0319 gene was calculated by combining the number of risk alleles of the 13 SNPs.

Interaction Analysis

Stratified regression analysis was performed to explore the interactions of 13 SNPs and the CGS with the parental education level. The standard multiple regression equation is as follows:

\(Y=B0+B1X1+B2X2+B3(X1 \times X2)+B4 \cdot Age+B5 \cdot Sex+E{\text{ (1)}}\)

where Y is the dependent variable (i.e., fluent reading); X₁ is the environmental variable (PE); X₂ is the genetic variable, X₁×X₂ is the product term of the gene‒environment interaction; B₁ and B₂ are the regression slopes of the main effects of environment X₁ and gene X₂, respectively; B₃ is the regression coefficient of their interaction term; B₀ is the intercept; and E is the random error.

Reparameterized Regression Model Tests

The reparametric equation [59] is as follows:

\(\begin{gathered} \hfill \\ Y=B0+B1(X1 - C)+B2((X1 - C) \times X2)+B3 \cdot Age+B4 \cdot Sex+E{\text{ (2)}} \hfill \\ \end{gathered}\)

Equation 2 is a 5-parameter equation (i.e., B0, B1, B2, B3, B4, C). C is the intersection of the predicted values of the environmental variables of the two groups; if the crossover of C and its confidence interval (CI) is within the range of values in the environment, the interaction is disordinal, reflecting the differential susceptibility model, and if it falls outside the range, the interaction is ordinal and conforms to the diathesis-stress model [59].

In this study, B1 is the slope of PE, and B2 is the slope of the interaction term. Point C is not fixed. If Point C is within the range of the parental education level, the interaction of G×E conforms to the differential susceptibility model. If Point C is fixed, a crossover point that falls at the maximum value of the environment variable is added (C = Max (PE)). At this point, the interaction of G×E is orderly and conforms to the diathesis-stress model. ANOVA, the Akaike information criterion (AIC) and the Bayesian information criterion (BIC) were used to evaluate the two models. For the AIC and BIC, the lower the value is, the higher the efficiency of the model.

Indirect Effects Analysis

A series of structural equation modeling (SEM) analyses were conducted by using Mplus 17.0 to explore whether RNA mediates the effect of cumulative genetic scores on fluent reading after controlling for age and sex. All the predictor measures were allowed to be related to each other. Indirect effects were tested using the 5000 bootstrap technique [60], and confidence intervals (95% CIs) that did not contain zero indicated significant indirect effects [61]. Moreover, we reported a variety of indices to reflect the model fit [62].

Correlation analysis

The 13 SNPs conformed to Hardy-Weinberg equilibrium (Supplementary Table S1). Table 1 presents the correlation among all variables for the total sample. The CGS of fluent reading (FR_CRS) was significantly correlated with fluent reading and digit rapid automatized naming. However, the CGS of character recognition (CR_CGS) was not correlated with any of the variables. Fluent reading and all of the rapid automatized naming tasks were significantly correlated. Moreover, there was no correlation between the environmental and genetic variables.

Table 1

Correlation matrix of Cumulative Genetic Scores, Rapid automatized naming, Fluent reading, Sex and Age.
	1	2	3	4	5	6	7	8	9	10	11
1.FR_GRS	1
2.CR_GRS	− .204**	1
3.Character Recognition	− .041	.027	1
4.Fluent Reading	.063*	.018	− .562**	1
5.Phonological Awareness	− .009	.010	.398**	− .331**	1
6.Morphological awareness	− .016	− .038	.325**	− .293**	.411**	1
7.RAN1	.056*	.043	− .331**	.594**	− .267**	− .216**	1
8.RAN2	.030	.029	− .276**	.436**	− .255**	− .194**	.564**	1
9.RAN3	.032	.046	− .227**	.436**	− .245**	− .206**	.476**	.553**	1
10.RAN4	.047	.054	− .222**	.379**	− .223**	− .182**	.454**	.527**	.617**	1
11.Parental Education	− .003	− .026	.121**	− .156**	.154**	.177**	− .087**	− .059*	− .080*	− .077*	1
Note:, p < 0.05;*, p < 0.01; RAN1: digit rapid automatized naming; RAN2: dice rapid automatized naming; RAN3: picture rapid automatized naming; RAN4: color rapid automatized naming.

Standard Exploratory Analysis

Standard regression equations were used to test the G×E effect of a single SNP, CR_CGS (Supplementary Table S2 and Table S3) and FR_CGS (Table 2). In fluent reading and character recognition, no single SNP reached significance in the interaction terms after Bonferroni correction (adjusted p = .0038). For all analyses, the PE level was significant in all regression main effects. In Table 2 in Model 1, before adding the interaction term, the main effect of the PE level was significant (B₁ = -5.21, p < .001), and the predicting effect of FR_CGS for fluent reading was not (B₂ = .54, p = .06). In Model 2, the G×E effect was significant (B₃ = .52, p = .03, R² = .09). Simple slope analysis showed (Fig. 1A) that with the increase in the PE level, the time of the low FR_CGS subgroup decreased significantly and was significantly lower than that of the high FR_CGS subgroup. However, the interaction between CR_CGS and the PE level did not predict character recognition (Fig. 1B).

Table 2

The interaction between the genetic risk score of KIAA0319 and parental education on Fluent Reading.
Parameter	Gene(G) and environment(E) main effects: Model 1	Main effects and G×E interaction: Model 2
B₀	252.76(10.39)	268.97(12.84)
B₁	-5.21(.73)	-10.26(2.47)
B₂	.54(.29)	-1.13(.83)
B₃	--	.52(.24)
B₄	− .83(.07)	− .83(.07)
B₅	-2.06(1.60)	-2.13(1.60)
R²	.084	.087
F	36.46	30.15
df	4, 1583	5, 1582
p	< .0001	< .0001
F vs. 1	--	6.308
df	--	1,1582
p	--	.032

Competing Test Selection

Reparameterized Eq. 2 was used to verify the interaction between the KIAA0319 gene and parental education level. Table 3 shows that Model 3 (i.e., the differential susceptibility model) had the best fitting effect on character recognition (Fig. 2). Compared with Model 4, Model 3 had an estimated parameter added, and the interpretation rate of R² increased significantly (△R² = .002, p = .008), so Model 4 was rejected. Furthermore, by comparing Model 3 and Model 4, the AIC and BIC of Model 4 were both larger than those of Model 3, and Model 3b was rejected. Overall, Model 3 had better fitting performance. In the G×E effect, FR_CGS and the parental education (PE) level fit the differential susceptibility model.

Table 3

Re-parameterized regression analysis
Linear×Linear Interaction
Parameter	Differential susceptibility Model3	Diathesis-Stress Model 4
B₀	246.69(9.79)	222.76(8.16)
B₁	-10.26(2.47)	-4.38(1.03)
B₂	.52(.24)	− .08(.07)
C	2.17(.74)	6.83(--)
B₃	− .82(.07)	− .82(.07)
B₄	-2.13(1.60)	-2.05(1.60)
R²	.0806	.0747
F	47.37	33.02
df	6,1581	5, 1582
p	< .0001	< .0001
F vs. 2	6.88	--
df	1,1581	--
p	.008	--
AIC	15511.69	15522.86
BIC	15538.54	15555.08
Note: AIC, Akaike information criterion; BIC, Bayesian information criterion.
Tabled values are parameter estimates, with their standard errors in parentheses.
F vs. 1, 3a and 3c, stands for an F test of the difference in R² for Model 2 versus Model 1, Model 3b versus Model 3a, and Model 3d versus Model 3c, respectively.
^a Parameter fixed at + 3 standard deviation; SE is not applicable, so is listed as (--).

Test Of The Mediation Model

A series of structural equation modeling (SEM) analyses were conducted to explore whether cognitive skills mediated the effect of the KIAA0319 gene on fluent reading and character recognition after controlling for sex and age (Fig. 3). All the predictor measures were allowed to correlate with each other. These indices of the modified models (Table 4 and Supplementary Table S4) provided a good fit to the data: the CFI (comparative fit index) = .999 > .90; the TLI (Tucker‒Lewis index) = .994 > .90; the RMSEA (root mean square error of approximation) = .022 < .80; andχ²/df = 2.14 < 3.

Table 4

Fitting index of cumulative genetic score on rapid naming mediation model.
	χ²	df	χ²/df	CFI	TLI	RMSEA
Model5	344.82	8	43.10	0.928	0.774	0.136
Modified Model6	8.55	4	2.14	0.999	0.994	0.022
Note: CFI = Comparative Fit Index; TLI = Tucker-Lewis Index; RMSEA = Root Mean Square Error of Approximation; CFA = Confirmatory Factor Analyses;

Using 5000 bootstrap analyses and 95% CIs, we only found that the significant mediation of digit RAN in fluent reading was [.004.466], and the confidence interval did not include 0 (Table 5) and explained 30.6% of the variance in fluent reading. According to the β values, the specific indirect effect pathways were positive. The result suggested that the higher the FR_CGS, the longer the RAN time, and correspondingly, the longer the fluent reading time. In Supplementary Table S5, the mediation model from CR_GRS to character recognition is shown, and no significant mediating effect was found.

Table 5

Specifific indirect effects of RAN, Phonological awarenes and MG from CGS to Fluent Reading in Modified Model 5 (standardized βs and SEs are reported).
	β	SE	95%CI*
FR_CGS → Digit RAN → Fluent Reading	.020	.010	[.004, .466]
FR_CGS →Dice RAN → Fluent Reading	.001	.001	[− .006, .070]
FR_CGS → Picture RAN → Fluent Reading	.005	.004	[-.019, .148]
FR_CGS → Color RAN → Fluent Reading	.002	.002	[.000, .077]
FR_CGS → Phonological awarenes → Fluent Reading	.003	.005	[-.007, .012]
FR_CGS → Morphological awareness → Fluent Reading	.006	.005	[-.003, .015]
* Significant coefficients are reported in bold.

The present study examined not only individual SNPs but also the cumulative effect of KIAA0319. The interaction between individual SNPs and the parental education level was not significant. However, the interaction between the CGS and the parental factor was significant, suggesting that KIAA0319 may affect children's reading abilities through multiple minor variations. At the same time, we also found that the CGS can affect children's fluent reading through a cognitive variable, i.e., digit RNA. This mediating effect mainly focuses on digit rapid automatized naming.

The SNPs we selected did not show significant interaction effects. A possible reason is the insufficient statistical power, since the number of samples verifying the interaction was more than the number of samples for which the effect was found alone [63]. The significant results of cumulative gene score might be the increased power of the cumulative effect, which was explored for the first time in this sample. The present study is the first to assess the cumulative effect of candidate genes on reading ability in Chinese children using the cumulative genetic risk score. As already noted, in most G×E work, one gene polymorphism is studied at a time. The CGS model accounts for more variance than individual SNPs, which, to a certain extent, reduced the problem of repeat-times to analyze individual polymorphisms [64].

A positive correlation indicated that individuals with a high CGS had worse fluent reading skills. Under the influence of the interaction, individuals with a low CGS were better at fluent reading in a positive environment than individuals with a high CGS. These results suggested that the more risk alleles an individual carries, the worse their performance, and KIAA0319 is theoretically a vulnerability gene. The GXE effect on fluent reading was found to be consistent with the differential susceptibility model by competing model analysis. This result indicated that “vulnerability genes” can be appropriately described as “plastic genes” because they make individuals more susceptible to environmental influences and thus exhibit better or worse behavior [49, 55]. In this study, the negative effects of KIAA0319 on reading ability were verified for the first time through reparameterized regression models. In other words, the more risk alleles an individual carries, the more vulnerable they are, potentially causing irreversible harm, and the environment has little effect on the individual. Conversely, individuals carrying fewer vulnerability alleles are more susceptible to environmental influences and thus perform well.

To some extent, the finding is consistent with the vantage sensitivity model due to the small value of the cross point. The notion is that salutary environments can moderate the influence of genetic variations on behaviors [48, 65] but not adverse environments [66]. Although there have been some findings of vantage sensitivity models in other domains, the study of cognitive ability has shown that the cumulative genetic scores of the COMT and DRD2 genes and the effect of father authoritarianism on creativity are consistent with the vantage sensitivity model [67], the evidence is lacking in the field of reading. Therefore, the results of the vantage sensitivity model need to be further validated.

Our study supports the mediating role of digit RAN in fluent reading [36, 68]. A total of 30.6% of the variance in fluent reading suggests that part of the inheritance of phenotypes is mediated through endophenotypes. The correlations between the CGS and RAN were consistent with previous results that KIAA0319 can affect RAN [12, 41]. However, in our study, we used β values derived from the phenotype of fluent reading, so the individual variant and cumulative genetic scores encoded do not provide an accurate estimate of the effect on RAN. This independent cumulative effect affects the final behavioral outcome through digit RAN, which means that, coincident with models of multiple deficits, the direct influence of genetic variation is limited [1, 69]. Contrary to our findings, previous studies did not provide sufficient evidence for RAN as an EP [31, 70], whereas in our study, RAN was found to mediate gene-phenotype associations. This might be due to different measurement methods of endophenotypes, different languages and fewer selected SNPs. Therefore, subsequent studies need to be repeated using standard RAN measurements. Unfortunately, we did not find any other significant mediating effects, and phonological awareness and morphological awareness need to be further investigated as endophenotypes of developmental dyslexia.

However, there were several limitations in the current study. First, we did not discuss the effects of individual SNPs on endophenotypes and collected previous fluent reading and character recognition results to test for mediating and moderating effects. Second, except for the digit RAN task, the other three rapid automatized naming tasks might have measurement biases, resulting in no significant results. Third, the current study did not investigate the overlapping effects of cognitive variables, and controlling for each variable could lead to opposite results. Finally, although the CGS adds statistical power, it should be duplicated in an independent cohort of more than 8,000 participants to verify the presence of the interaction.

KIAA0139, a candidate gene for reading ability, is a risk factor, not a protective factor. The more risk alleles a person carries, the worse their reading fluency is. The finding that KIAA0319 interacts with the parental education level suggested that environmental variables can modulate the effects of an individual's genes on their behaviors [14]. Through the sequencing of the mediator variables shown, candidate genes can drive the function of cognitive intermediate phenotypes and thus influence reading outcomes. This study, from a genetic perspective, suggests that rapid automatized naming interventions can help improve reading ability.

Ethics approval and consent to participate

All experimental procedures have been authorized by the Shaanxi Normal University and written informed consent was obtained for all participant children’s parents.

Consent for publication

Not applicable.

Availability of data and materials

Requests for the R code can be made through the first author of this publication.

Competing interests

The research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding

This study was supported by Fund for Humanities and Social Sciences Research of the Ministry of Education (19YJC190023), the China Postdoctoral Science Foundation funding project (2019M663924XB), Natural Science Foundation of Shaanxi Province (2021JQ-309), and Planning Subject for the 14th Five Year Plan of Shaanxi Education Sciences (SGH21Y0040).

Authors' contributions

QY, CC, and ZW performed the experiments and data analysis and assisted in writing the manuscript. QY and CC performed the experiments and analyzed the data. ZW and QY designed the study and wrote the manuscript. All authors approved the final version of the manuscript.

Pennington BF. From single to multiple deficit models of developmental disorders. Cognition. 2006;101(2):385–413.
Shaywitz SE, Shaywitz BA. Dyslexia (Specific Reading Disability). Biol Psychiatry. 2005;57(11):1301–9.
Schmitz J, Kumsta R, Moser D, Güntürkün O, Ocklenburg S. KIAA0319 promoter DNA methylation predicts dichotic listening performance in forced-attention conditions. Behav Brain Res. 2018;337:1–7.
Szalkowski CE, Fiondella CF, Truong DT, Rosen GD, LoTurco JJ, Fitch RH. The effects of Kiaa0319 knockdown on cortical and subcortical anatomy in male rats. Int J Dev Neurosci. 2013;31(2):116–22.
Velayos-Baeza A, Levecque C, Kobayashi K, Holloway ZG, Monaco AP. The Dyslexia-associated KIAA0319 Protein Undergoes Proteolytic Processing with γ-Secretase-independent Intramembrane Cleavage*. J Biol Chem. 2010;285(51):40148–62.
Peschansky VJ, Burbridge TJ, Volz AJ, Fiondella C, Wissner-Gross Z, Galaburda AM, Turco JJL, Rosen GD. The Effect of Variation in Expression of the Candidate Dyslexia Susceptibility Gene Homolog Kiaa0319 on Neuronal Migration and Dendritic Morphology in the Rat. Cereb Cortex. 2009;20(4):884–97.
Poon M-W, Tsang W-H, Chan S-O, Li H-M, Ng H-K, Waye MM-Y. Dyslexia-Associated Kiaa0319-Like Protein Interacts with Axon Guidance Receptor Nogo Receptor 1. Cell Mol Neurobiol. 2011;31(1):27–35.
Gostic M, Martinelli A, Tucker C, Yang Z, Gasparoli F, Ewart J-Y, Dholakia K, Sillar KT, Tello JA, Paracchini S. The dyslexia susceptibility KIAA0319 gene shows a specific expression pattern during zebrafish development supporting a role beyond neuronal migration. J Comp Neurol. 2019;527(16):2634–43.
Francks C, Paracchini S, Smith SD, Richardson AJ, Scerri TS, Cardon LR, Marlow AJ, MacPhie IL, Walter J, Pennington BF, et al. A 77-Kilobase Region of Chromosome 6p22.2 Is Associated with Dyslexia in Families From the United Kingdom and From the United States. Am J Hum Genet. 2004;75(6):1046–58.
Paracchini D, Phil S, Steer CD, Buckingham L-L, Morris AP, Ring S, Scerri D, Phil T, Stein J, Pembrey ME, Ragoussis J, Golding J. Association of the KIAA0319 dyslexia susceptibility gene with reading skills in the general population. Am J Psychiatry. 2008;165(12):1576–84.
Scerri TS, Morris AP, Buckingham L-L, Newbury DF, Miller LL, Monaco AP, Bishop DVM, Paracchini S. DCDC2, KIAA0319 and CMIP Are Associated with Reading-Related Traits. Biol Psychiatry. 2011;70(3):237–45.
Carrion-Castillo A, Maassen B, Franke B, Heister A, Naber M, van der Leij A, Francks C, Fisher SE. Association analysis of dyslexia candidate genes in a Dutch longitudinal sample. Eur J Hum Genet. 2017;25(4):452–60.
Becker J, Czamara D, Scerri TS, Ramus F, Csépe V, Talcott JB, Stein J, Morris A, Ludwig KU, Hoffmann P, et al. Genetic analysis of dyslexia candidate genes in the European cross-linguistic NeuroDys cohort. Eur J Hum Genet. 2014;22(5):675–80.
Mascheretti S, Riva V, Giorda R, Beri S, Lanzoni LFE, Cellino MR, Marino C. KIAA0319 and ROBO1: evidence on association with reading and pleiotropic effects on language and mathematics abilities in developmental dyslexia. J Hum Genet. 2014;59(4):189–97.
Manuck SB, McCaffery JM. Gene-Environment Interaction. Ann Rev Psychol. 2014;65(1):41–70.
Irving I, Gottesman PhD, Hon FRCPsych, Gould TD. M.D.: The Endophenotype Concept in Psychiatry: Etymology and Strategic Intentions. Am J Psychiatry. 2003;160(4):636–45.
Kendler KS, Neale MC. Endophenotype: a conceptual analysis. Mol Psychiatry. 2010;15(8):789–97.
Flint J, Timpson N, Munafò M. Assessing the utility of intermediate phenotypes for genetic mapping of psychiatric disease. Trends Neurosci. 2014;37(12):733–41.
Braff DL. The importance of endophenotypes in schizophrenia research. Schizophr Res. 2015;163(1):1–8.
Szatmari P, Maziade M, Zwaigenbaum L, Mérette C, Roy M-A, Joober R, Palmour R. Informative phenotypes for genetic studies of psychiatric disorders. Am J Med Genet Part B: Neuropsychiatric Genet. 2007;144B(5):581–8.
Jones MW, Ashby J, Branigan HP. Dyslexia and fluency: Parafoveal and foveal influences on rapid automatized naming. J Exp Psychol Hum Percept Perform. 2013;39:554–67.
Liao C-H, Deng C, Hamilton J, Lee CS-C, Wei W, Georgiou GK. The role of rapid naming in reading development and dyslexia in Chinese. J Exp Child Psychol. 2015;130:106–22.
Bowey JA, McGuigan M, Ruschena A. On the association between serial naming speed for letters and digits and word-reading skill: towards a developmental account. J Res Reading. 2005;28(4):400–22.
Pan J, McBride-Chang C, Shu H, Liu H, Zhang Y, Li H. What is in the naming? A 5-year longitudinal study of early rapid naming and phonological sensitivity in relation to subsequent reading skills in both native Chinese and English as a second language. J Educ Psychol. 2011;103:897–908.
Savage R, Frederickson N. Evidence of a highly specific relationship between rapid automatic naming of digits and text-reading speed. Brain Lang. 2005;93(2):152–9.
Truong DT, Adams AK, Paniagua S, Frijters JC, Boada R, Hill DE, Lovett MW, Mahone EM, Willcutt EG, Wolf M, et al. Multivariate genome-wide association study of rapid automatised naming and rapid alternating stimulus in Hispanic American and African–American youth. J Med Genet. 2019;56(8):557.
Gialluisi A, Andlauer TFM, Mirza-Schreiber N, Moll K, Becker J, Hoffmann P, Ludwig KU, Czamara D, St Pourcain B, Brandler W, et al. Genome-wide association scan identifies new variants associated with a cognitive predictor of dyslexia. Translational Psychiatry. 2019;9(1):77.
Andreola C, Mascheretti S, Belotti R, Ogliari A, Marino C, Battaglia M, Scaini S. The heritability of reading and reading-related neurocognitive components: A multi-level meta-analysis. Neurosci Biobehavioral Reviews. 2021;121:175–200.
Olson RK, Hulslander J, Christopher M, Keenan JM, Wadsworth SJ, Willcutt EG, Pennington BF, DeFries JC. Genetic and environmental influences on writing and their relations to language and reading. Ann Dyslexia. 2013;63(1):25–43.
Byrne B, Samuelsson S, Wadsworth S, Hulslander J, Corley R, DeFries JC, Quain P, Willcutt EG, Olson RK. Longitudinal twin study of early literacy development: Preschool through Grade 1. Read Writ. 2007;20(1):77–102.
Mascheretti S, Riva V, Feng B, Trezzi V, Andreola C, Giorda R, Villa M, Dionne G, Gori S, Marino C, et al. The Mediation Role of Dynamic Multisensory Processing Using Molecular Genetic Data in Dyslexia. Brain Sci. 2020;10(12):993.
Glahn DC, Knowles EEM, McKay DR, Sprooten E, Raventós H, Blangero J, Gottesman II, Almasy L. Arguments for the sake of endophenotypes: Examining common misconceptions about the use of endophenotypes in psychiatric genetics. Am J Med Genet Part B: Neuropsychiatric Genet. 2014;165(2):122–30.
Kirby JR, Deacon SH, Bowers PN, Izenberg L, Wade-Woolley L, Parrila R. Children’s morphological awareness and reading ability. Read Writ. 2012;25(2):389–410.
Ku Y-M, Anderson RC. Development of morphological awareness in Chinese and English. Read Writ. 2003;16(5):399–422.
Pan J, Song S, Su M, McBride C, Liu H, Zhang Y, Li H, Shu H. On the relationship between phonological awareness, morphological awareness and Chinese literacy skills: evidence from an 8-year longitudinal study. Dev Sci. 2016;19(6):982–91.
Tighe EL, Schatschneider C. Examining the Relationships of Component Reading Skills to Reading Comprehension in Struggling Adult Readers:A Meta-Analysis. J Learn Disabil. 2016;49(4):395–409.
Betjemann RS, Keenan JM, Olson RK, DeFries JC. Choice of Reading Comprehension Test Influences the Outcomes of Genetic Analyses. Sci Stud Read. 2011;15(4):363–82.
Gayán J, Olson RK. Genetic and environmental influences on individual differences in printed word recognition. J Exp Child Psychol. 2003;84(2):97–123.
Xie Q, Zheng M, Ho CS-H, McBride C, Fong FLW, Wong SWL, Chow BW-Y. Exploring the Genetic and Environmental Etiologies of Phonological Awareness, Morphological Awareness, and Vocabulary Among Chinese–English Bilingual Children: The Moderating Role of Second Language Instruction. Behav Genet. 2022;52(2):108–22.
Lim CK-P, Wong AM-B, Ho CS-H, Waye MM-Y. A common haplotype of KIAA0319 contributes to the phonological awareness skill in Chinese children. Behav Brain Funct. 2014;10(1):23.
Venkatesh SK, Siddaiah A, Padakannaya P, Ramachandra NB. Analysis of genetic variants of dyslexia candidate genes KIAA0319 and DCDC2 in Indian population. J Hum Genet. 2013;58(8):531–8.
D'Souza S, Backhouse-Smith A, Thompson JMD, Slykerman R, Marlow G, Wall C, Murphy R, Ferguson LR, Mitchell EA, Waldie KE. Associations Between the KIAA0319 Dyslexia Susceptibility Gene Variants, Antenatal Maternal Stress, and Reading Ability in a Longitudinal Birth Cohort. Dyslexia. 2016;22(4):379–93.
Plak RD, Merkelbach I, Kegel CAT, van Ijzendoorn MH, Bus AG. Brief computer interventions enhance emergent academic skills in susceptible children: A gene-by-environment experiment. Learn Instruction. 2016;45:1–8.
Su M, Wang J, Maurer U, Zhang Y, Li J, McBride C, Tardif T, Liu Y, Shu H. Gene–environment interaction on neural mechanisms of orthographic processing in Chinese children. J Neurolinguistics. 2015;33:172–86.
Su M, Li P, Zhou W, Shu H. Effects of socioeconomic status in predicting reading outcomes for children: The mediation of spoken language network. Brain Cogn. 2021;147:105655.
Su M, Peyre H, Song S, McBride C, Tardif T, Li H, Zhang Y, Liang W, Zhang Z, Ramus F, et al: The influence of early linguistic skills and family factors on literacy acquisition in Chinese children: Follow-up from age 3 to age 11. Learning and Instruction 2017, 49:54–63.
Hyde LW, Bogdan R, Hariri AR. Understanding risk for psychopathology through imaging gene–environment interactions. Trends Cogn Sci. 2011;15(9):417–27.
Manuck SB, McCaffery JM. Genetics of stress: Gene–stress correlation and interaction. In: Handbook of behavioral medicine. Springer; 2010. pp. 455–78.
Belsky J, Pluess M. Beyond diathesis stress: Differential susceptibility to environmental influences. Psychol Bull. 2009;135:885–908.
Belsky J, Jonassaint C, Pluess M, Stanton M, Brummett B, Williams R. Vulnerability genes or plasticity genes? Mol Psychiatry. 2009;14(8):746–54.
Enge S, Sach M, Reif A, Lesch K-P, Miller R, Fleischhauer M. Cumulative Dopamine Genetic Score predicts behavioral and electrophysiological correlates of response inhibition via interactions with task demand. Cogn Affect Behav Neurosci. 2020;20(1):59–75.
Pearson R, Palmer RHC, Brick LA, McGeary JE, Knopik VS, Beevers CG. Additive genetic contribution to symptom dimensions in major depressive disorder. J Abnorm Psychol. 2016;125:495–501.
Steiger H, Thaler L, Gauvin L, Joober R, Labbe A, Israel M, Kucer A. Epistatic interactions involving DRD2, DRD4, and COMT polymorphisms and risk of substance abuse in women with binge-purge eating disturbances. J Psychiatr Res. 2016;77:8–14.
Walton E, Turner J, Gollub RL, Manoach DS, Yendiki A, Ho B-C, Sponheim SR, Calhoun VD, Ehrlich S. Cumulative Genetic Risk and Prefrontal Activity in Patients With Schizophrenia. Schizophr Bull. 2012;39(3):703–11.
Belsky J, Beaver KM. Cumulative-genetic plasticity, parenting and adolescent self-regulation. J Child Psychol Psychiatry. 2011;52(5):619–26.
Shu H, McBride-Chang C, Wu S, Liu H. Understanding chinese developmental dyslexia: Morphological awareness as a core cognitive construct. J Educ Psychol. 2006;98:122–33.
Li H, Shu H, McBride-Chang C, Liu H, Peng H. Chinese children's character recognition: Visuo-orthographic, phonological processing and morphological skills. J Res Reading. 2012;35(3):287–307.
Doust C, Fontanillas P, Eising E, Gordon SD, Wang Z, Alagöz G, Molz B, Pourcain BS, Francks C, Marioni RE, et al: Discovery of 42 Genome-Wide Significant Loci Associated with Dyslexia. medRxiv 2021:2021.2008.2020.21262334.
Widaman KF, Helm JL, Castro-Schilo L, Pluess M, Stallings MC, Belsky J. Distinguishing ordinal and disordinal interactions. Psychol Methods. 2012;17:615–22.
Fritz MS, MacKinnon DP. Required Sample Size to Detect the Mediated Effect. Psychol Sci. 2007;18(3):233–9.
Tofighi D, MacKinnon DP. RMediation: An R package for mediation analysis confidence intervals. Behav Res Methods. 2011;43(3):692–700.
Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behav Res Methods. 2008;40(3):879–91.
Thomas D. Gene–environment-wide association studies: emerging approaches. Nat Rev Genet. 2010;11(4):259–72.
Docherty SJ, Kovas Y, Plomin R. Gene-Environment Interaction in the Etiology of Mathematical Ability Using SNP Sets. Behav Genet. 2011;41(1):141–54.
Sweitzer MM, Halder I, Flory JD, Craig AE, Gianaros PJ, Ferrell RE, Manuck SB. Polymorphic variation in the dopamine D4 receptor predicts delay discounting as a function of childhood socioeconomic status: evidence for differential susceptibility. Soc Cognit Affect Neurosci. 2012;8(5):499–508.
Pluess M, Belsky J. Vantage sensitivity: Individual differences in response to positive experiences. Psychol Bull. 2013;139:901–16.
Si S, Su Y, Zhang S, Zhang J. Genetic susceptibility to parenting style: DRD2 and COMT influence creativity. NeuroImage. 2020;213:116681.
Norton ES, Wolf M. Rapid Automatized Naming (RAN) and Reading Fluency: Implications for Understanding and Treatment of Reading Disabilities. Ann Rev Psychol. 2012;63(1):427–52.
McGrath LM, Peterson RL, Pennington BF. The Multiple Deficit Model: Progress, Problems, and Prospects. Sci Stud Read. 2020;24(1):7–13.
Mascheretti S, Gori S, Trezzi V, Ruffino M, Facoetti A, Marino C. Visual motion and rapid auditory processing are solid endophenotypes of developmental dyslexia. Genes Brain and Behavior. 2018;17(1):70–81.

No competing interests reported.

supplement.docx

Download PDF

Journal Publication

published 01 Jun, 2023

Read the published version in Behavioral and Brain Functions →

Editorial decision: Major revision
30 Nov, 2022
Reviews received at journal
24 Oct, 2022
Reviewers agreed at journal
12 Oct, 2022
Reviewers invited by journal
06 Oct, 2022
Editor assigned by journal
27 Sep, 2022
Submission checks completed at journal
27 Sep, 2022
First submitted to journal
24 Sep, 2022

You are reading this latest preprint version

Cumulative genetic score of KIAA0319 affects reading fluency through parental education levels and rapid automatized naming

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Materials And Methods

Participants

Genotyping And Snp Selection

Phenotypes

Parental Education (Pe) Levels

Data analysis

SNP coding and cumulative genetic scores (CGSs)

Interaction Analysis

Reparameterized Regression Model Tests

Indirect Effects Analysis

Results

Correlation analysis

Standard Exploratory Analysis

Competing Test Selection

Test Of The Mediation Model

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1