Autism spectrum disorder (ASD) is a diverse neurodevelopmental condition with a significant genetic basis. Machine learning (ML) and deep learning (DL) models are emerging as powerful tools for identifying potential ASD risk genes and biomarkers to improve diagnostic accuracy. This narrative systematic review synthesizes evidence on the use of ML and DL models for ASD prediction and diagnosis, focusing on studies that utilized genetic variants and biomarkers. We conducted a systematic search across six databases [Web of Science, PubMed, Science Direct, Semantic Scholar, Institute of Electrical and Electronics Engineers (IEEE Xplore), and Association for Computing Machinery Digital Library (ACM-DL)] for articles published between 2009 and 2024, yielding 483 initial records. Multiple inclusion criteria were applied, and only peer-reviewed articles that used artificial intelligence, ML, or DL to predict ASD based on genetic biomarkers were included. After deduplication and applying inclusion and exclusion criteria, 14 studies were included in this review. These studies focused on genetic variants based on whole-genome sequencing, whole-exome sequencing, RNA sequencing, single-nucleus RNA sequencing, mRNA expression, long non-coding RNA, or single nucleotide polymorphism data. Our findings indicate that various ML models achieved robust performance, while DL architectures such as DeepASDPred outperformed ML models, with an accuracy of 93.8%. Most of the studies used datasets such as the Simons Foundation Autism Research Initiative (SFARI), the Simons Simplex Collection (SSC), the Autism Genetic Resource Exchange (AGRE), Atlas of the Developing Human Brain (BrainSpan), and the Simons Foundation Powering Autism Research for Knowledge (SPARK). The reviewed studies demonstrate that ML can help identify genetic risk factors and biomarkers that contribute to ASD. However, challenges remain, including data heterogeneity, small sample sizes, and the need for greater model interpretability. Addressing these challenges is crucial for translating ML models into clinically useful tools. In conclusion, DL substantially enhances ASD prediction, leveraging genetic data to identify novel candidate genes and support biological discovery of cell type-specific gene expression and roles of non-coding mutations in ASD.
KIF1A-associated neurological diseases (KANDs) are a group of inherited conditions caused by changes in the microtubule (MT) motor protein KIF1A as a result of KIF1A gene mutations. Anterograde transport of membrane organelles is facilitated by the kinesin family protein encoded by the MT-based motor gene KIF1A. Variations in the KIF1A gene, which primarily affect the motor domain, disrupt its ability to transport synaptic vesicles containing synaptophysin and synaptotagmin leading to various neurological pathologies such as hereditary sensory neuropathy, autosomal dominant and recessive forms of spastic paraplegia, and different neurological conditions. These mutations are frequently misdiagnosed because they result from spontaneous, non-inherited genomic alterations. Whole-exome sequencing (WES), a cutting-edge method, assists neurologists in diagnosing the illness and in planning and choosing the best course of action. These conditions are simple to be identified in pediatric and have a life expectancy of 5–7 years. There is presently no permanent treatment for these illnesses, and researchers have not yet discovered a medicine to treat them. Scientists have more hope in gene therapy since it can be used to cure diseases brought on by mutations. In this review article, we discussed some of the experimental gene therapy methods, including gene replacement, gene knockdown, symptomatic gene therapy, and cell suicide gene therapy. It also covered its clinical symptoms, pathogenesis, current diagnostics, therapy, and research advances currently occurring in the field of KAND-related disorders. This review also explained the impact that gene therapy can be designed in this direction and afford the remarkable benefits to the patients and society.
kinesin motor proteinmicrotubule motor proteinsanterograde transportsynaptic vesicle transport
Abstract One of the most common inborn errors in fatty acid β oxidation (FAO) is a very long-chain acyl-coenzyme A dehydrogenase (VLCAD) deficiency. It is autosomal recessive. The enzyme used in the first phase of FAO is VLCAD. The enzyme is responsible for β oxidation spiral pathway's initial step, the dehydrogenation process of long-chain fatty acyl-CoA. The phenotypes include hypoglycemia, hepatomegaly, cardiomyopathy, and occasionally abrupt mortality. Most VLCAD deficiencies in newborns are now detected during the neonatal period due to the development of newborn screening programs. Mitochondrial DNA depletion syndromes (MTDPS) are one of the rarest metabolic disorders. It is an autosomal recessive disease caused by defects in genes necessary for the maintenance of mitochondrial DNA (mtDNA). One of these FBXL4 (F-box and leucine-rich repeat protein 4) variants causes encephalomyopathic mtDNA depletion syndrome 13 (MTDPS13), which presents as a failure to thrive, severe global developmental delay, hypotonia, early infantile onset of encephalopathy, and lactic acidosis. We report here the case of a Saudi infant born to consanguineous parents who presented to us with severe failure to thrive, profound neurodevelopmental delays, and facial dysmorphic features. Whole-exome sequencing (WES) showed the infants had MTDPS13. The FBXL4 variant c.1698A > G p. (Ile566Met) has previously been described as a disease that causes developmental delay and lactic acidosis, and another variant has also been detected in the patient. The ACADVL variant c.134C > A p. (Ser45*) has previously been described to cause VLCAD deficiency. A comprehensive literature review showed our patient to be the first case of MTDPS13 and VLCAD reported to date worldwide.
very long-chain acyl-CoA dehydrogenase deficiencyfatty acid beta-oxidation disordersmitochondrial DNA depletion syndromeconsanguinity