Identifying interactive biological pathways associated with reading disability

Abstract

Introduction Past research has suggested that reading disability is a complex disorder involving genetic and environment contributions, as well as gene–gene and gene–environment interaction, but to date little is known about the underlying mechanisms. Method Using the Avon Longitudinal Study of Parents and Children, we assessed the contributions of genetic, demographic, and environmental variables on case–control status using machine learning. We investigated the functional interactions between genes using pathway and network analysis. Results Our results support a systems approach to studying the etiology of reading disability with many genes (e.g., RAPGEF2 , KIAA0319 , DLC1 ) and biological pathways (e.g., neuron migration, positive regulation of dendrite regulation, nervous system development) interacting with each other. We found that single nucleotide variants within genes often had opposite effects and that enriched biological pathways were mediated by neuron migration. We also identified behavioral (i.e., receptive language, nonverbal intelligence, and vocabulary), demographic (i.e., mother's highest education), and environmental (i.e., birthweight) factors that influenced case–control status when accounting for genetic information. Discussion The behavioral and demographic factors were suggested to be protective against reading disability status, while birthweight conveyed risk. We provided supporting evidence that reading disability has a complex biological and environmental etiology and that there may be a shared genetic and neurobiological architecture for reading (dis)ability.

Genetic and Demographic Predictors of Latent Reading Ability in Two Cohorts

“…Variants for CYP19A1:rs10046 and DNAAF4:rs77641139 can affect regulatory elements of these genes as they can alter the mRNA either in the 3' UTR or through nonsense mediated decay, providing a possible method by which they influence our reading ability phenotype. SNPs from RAPGEF2 and NEDD4L have also been previously associated with reading disability 19,74 and related cognitive skills (e.g.,…”

“…reading 16 ) or correlated with reading (e.g., multivariate rapid automatized naming/rapid alternating stimulus 17 ). Recent investigations involving reading disability and related tasks have provided evidence for polygenicity, gene-gene interactions, 18 and functional biological pathways 19,20 . For example, KIAA0319/TTRAP and DYX1C1 interact with GRIN2B in children with a reading disability when performing a short-term memory task 21 and DCDC2/KIAA0319 interact to diminish single word reading, nonword reading, spelling, phoneme deletion, and comprehension 22 .…”

“…By combining machine learning (i.e., statistical models that learn from data) and confirmatory analytical methods, we can further our understanding of the genetics of reading ability and contribute to the refinement of the field's hypotheses. We have previously developed methods for combining statistical and machine learning approaches with biological domain knowledge to study the association between genetic and environmental factors and disease [31][32][33] , and have applied them to study various disorders 19,[34][35][36][37][38][39][40][41][42][43] , including those involving reading ability 19 .…”

“…Past research has established genetic contributions to reading disability (i.e., difficulties learning to read which cannot be explained by neurological or sensorial conditions, including dyslexia and other subtypes) 13–15 and performance on quantitative traits used to determine reading disability status (e.g., nonword reading 16 ) or correlated with reading (e.g., multivariate rapid automatized naming/rapid alternating stimulus 17 ). Recent investigations involving reading disability and related tasks have provided evidence for polygenicity, gene-gene interactions, 18 and functional biological pathways 19, 20 . For example, KIAA0319/TTRAP and DYX1C1 interact with GRIN2B in children with a reading disability when performing a short-term memory task 21 and DCDC2/KIAA0319 interact to diminish single word reading, nonword reading, spelling, phoneme deletion, and comprehension 22 .…”

“…There is limited knowledge of how environmental and demographic factors interact with genetic factors because to date only three studies have examined gene-environment interactions in reading 25, 29, 30 . Due to constraints imposed by research design and statistical analysis, few studies have integrated genetic, environmental, and demographic data within the same analysis 19 . Beyond statistical constraints, another limiting factor in understanding the interactions between genes and environmental-demographic features is the demographic similarity between the most frequently used cohorts.…”