Application of Single-Nucleotide Polymorphisms in the Diagnosis of Autism Spectrum Disorders: A Preliminary Study with Artificial Neural Networks

Ghafouri‐Fard, Soudeh; Taheri, Mohammad; Omrani, Mir Davood; Daaee, Amir; Mohammad‐Rahimi, Hossein; Kazazi, Hosein

doi:10.1007/s12031-019-01311-1

“…Among the assessed SNPs, rs6782011 of GRM7 was recognized as the one of most signi cant risk factors related to ASD [23]. The discrepant nding in our study may be due to differences in phenotypic or genetic characteristics.…”

Section: Discussioncontrasting

confidence: 68%

A Study of Genetic Variants in Genes of Glutamate Signaling and Risk of Childhood Autism Spectrum Disorder

Liu

¹

,

Yu

²

,

Hu

³

et al. 2020

Preprint

0

View full text Add to dashboard Cite

Background: Dysfunction of glutamate signaling has been implicated in the etiology of autism spectrum disorder (ASD). This case-control study was to examine the association between childhood ASD and single nucleotide polymorphisms (SNPs) in genes of the glutamate signaling pathway in a Chinese Han population.Methods: A total of 12 SNPs in the SLC1A1, SLC25A12, GRM7 and GRM8 genes were examined. The Children Autism Rating Scale (CARS) was applied to evaluate the severity of the disease. The relationship between SNPs and the risk of ASD or the severity of the disease was determined by logistic regression.Results: The T allele of rs2292813 in the SLC25A12 gene was significantly associated with an increased risk of ASD (odds ratio (OD) =1.7, 95% confidence interval (CI): 1.1-2.6, P=0.0107). Other examined SNPs were not associated with the risk of ASD. None of the SNPs examined were associated with the severity of ASD.Conclusions: Our findings support the involvement of SNPs in the SLC25A12 gene, but not SNPs in the SLC1A1, GRM7 and DRM8 genes, in the development of childhood ASD in the Chinese Han population.

show abstract

“…Our best accuracy (85.18%) was only 2.02% behind the best classifier found in the studies above, which uses EEG brain features (Duffy & Als, 2012). Indeed, our study surpass in accuracy the study mentioned above in the genetic field by about 11.51% (Ghafouri-Fard et al, 2019).…”

Section: Discussioncontrasting

confidence: 55%

“…First, we used a sample size of 2.636 participants whereas in other studies aiming to diagnostic classification in ASD the maximum number of samples used was around 1000 (Abraham et al, 2017;Duffy & Als, 2012;Ghafouri-Fard et al, 2019;Ghiassian, Greiner, Jin, & Brown, 2016;Heinsfeld, Franco, Craddock, Buchweitz, & Meneguzzi, 2018;Parikh, Li, & He, 2019;Subbaraju, Suresh, Sundaram, & Narasimhan, 2017). Our best accuracy (85.18%) was only 2.02% behind the best classifier found in the studies above, which uses EEG brain features (Duffy & Als, 2012).…”

Section: Discussionmentioning

confidence: 93%

“…We detected that of the 57 studies, only one focus on genetic features to predict ASD diagnosis. On this particular study (Ghafouri-Fard et al, 2019), 487 ASD patients and 455 healthy individuals were used to build a ML classifier which relied on single nucleotide polymorphism (SNP) data to predict the diagnosis. The accuracy, sensitivity and specificity achieved by their model were 73.67%, 82.75% and 63.95%, respectively.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Dopaminergic Gene Dosage in Autism versus Developmental Delay: From Complex Networks to Machine Learning approaches

Santos

¹

,

Caramelo

²

,

Melo

³

2020

Preprint

0

View full text Add to dashboard Cite

we studied genetic alterations related to dopamine processes that could impact brain development and function of 2636 participants. On average, from a genetic sample we were able to correctly separate autism from developmental delay with an accuracy of 85%. AbstractThe neural basis of behavioural changes in Autism Spectrum Disorders (ASD) remains a controversial issue. One factor contributing to this challenge is the phenotypic heterogeneity observed in ASD, which suggests that several different system disruptions may contribute to diverse patterns of impairment between and within study samples. Here, we took a retrospective approach, using SFARI data to study ASD by focusing on participants with genetic imbalances targeting the dopaminergic system. Using complex network analysis, we investigated the relations between participants, Gene Ontology (GO) and gene dosage related to dopaminergic neurotransmission from a polygenic point of view. We converted network analysis into a machine learning binary classification problem to differentiate ASD diagnosed participants from DD (developmental delay) diagnosed participants. Using 1846 participants to train a Random Forest algorithm, our best classifier achieved on average a diagnosis predicting accuracy of 85.18% (sd 1.11%) on a test sample of 790 participants using gene dosage features. In addition, we observed that if the classifier uses GO features it was also able to infer a correct response based on the previous examples because it is tied to a set of biological process, molecular functions and cellular components relevant to the problem. This yields a less variable and more compact set of features when comparing with gene dosage classifiers. Other facets of knowledge-based systems approaches addressing ASD through network analysis and machine learning, providing an interesting avenue of research for the future, are presented through the study.

show abstract

“…The correlation between SNPs in SLC1A1, SLC25A12, GRM7 and GRM8 genes has been reported in previous studies with inconsistent results [18,19] [20][21][22][23]. It is noted that most studies investigated these genes individually.…”

Section: Introductionmentioning

confidence: 92%

A Study of Genetic Variants in Genes of Glutamate Signaling and Risk of Childhood Autism Spectrum Disorder

Liu

¹

,

Yu

²

,

Hu

³

et al. 2020

Preprint

0

View full text Add to dashboard Cite

Background: Dysfunction of glutamate signaling has been implicated in the etiology of autism spectrum disorder (ASD). This case-control study was to examine the association between childhood ASD and single nucleotide polymorphisms (SNPs) in genes of the glutamate signaling pathway in a Chinese Han population.Methods: A total of 12 SNPs in the SLC1A1, SLC25A12, GRM7 and GRM8 genes were examined. The Children Autism Rating Scale (CARS) was applied to evaluate the severity of the disease. The relationship between SNPs and the risk of ASD or the severity of the disease was determined by logistic regression.Results: The T allele of rs2292813 in the SLC25A12 gene was significantly associated with an increased risk of ASD (odds ratio (OD) =1.7, 95% confidence interval (CI): 1.1-2.6, P=0.0107). Other examined SNPs were not associated with the risk of ASD. None of the SNPs examined were associated with the severity of ASD.Conclusions: Our findings support the involvement of SNPs in the SLC25A12 gene, but not SNPs in the SLC1A1, GRM7 and DRM8 genes, in the development of childhood ASD in the Chinese Han population.

show abstract

Application of Single-Nucleotide Polymorphisms in the Diagnosis of Autism Spectrum Disorders: A Preliminary Study with Artificial Neural Networks

Cited by 32 publications

References 20 publications

A Study of Genetic Variants in Genes of Glutamate Signaling and Risk of Childhood Autism Spectrum Disorder

A Study of Genetic Variants in Genes of Glutamate Signaling and Risk of Childhood Autism Spectrum Disorder

Dopaminergic Gene Dosage in Autism versus Developmental Delay: From Complex Networks to Machine Learning approaches

A Study of Genetic Variants in Genes of Glutamate Signaling and Risk of Childhood Autism Spectrum Disorder

Contact Info

Product

Resources

About