Front. Neurol. Frontiers in Neurology Front. Neurol. 1664-2295 Frontiers Media S.A. 10.3389/fneur.2026.1742186 Original Research Unmasking genetic etiologies in neurodevelopmental disorders characterized by Cerebral Palsy: insights from integrative genomic approaches Yigit Ayca 1 2 3 4 † Methodology Visualization Conceptualization Validation Writing – original draft Formal analysis Investigation Writing – review & editing Data curation Akgun-Dogan Ozlem 4 5 6 7 † Writing – original draft Investigation Conceptualization Validation Supervision Project administration Methodology Writing – review & editing Ozkeserli Zeynep 8 Writing – review & editing Data curation Visualization Formal analysis Writing – original draft Bayram Akcapınar Gunseli 8 Conceptualization Writing – review & editing Investigation Supervision Writing – original draft Funding acquisition Data curation Project administration Ayta Semih 9 Validation Writing – review & editing Investigation Resources Writing – original draft Gencpinar Pinar 10 Writing – original draft Resources Validation Writing – review & editing Investigation Maras Genc Hulya 11 Validation Writing – review & editing Writing – original draft Investigation Resources Kutlubay Busra 11 Investigation Writing – original draft Resources Validation Writing – review & editing Kara Bulent 12 Writing – review & editing Writing – original draft Resources Validation Investigation Sozen Hatice Gulhan 13 Investigation Writing – original draft Writing – review & editing Resources Validation Agaoglu Nihat Bugra 14 15 Writing – review & editing Writing – original draft Investigation Resources Validation Ozdemir Ozkan 4 16 Software Investigation Writing – original draft Methodology Writing – review & editing Supervision Bilguvar Kaya 4 5 7 17 * Project administration Validation Methodology Writing – original draft Supervision Formal analysis Resources Writing – review & editing Conceptualization Investigation Funding acquisition Ozbek Ugur 1 3 4 * Supervision Writing – review & editing Project administration Funding acquisition Investigation Writing – original draft Resources Conceptualization Genome Studies, Graduate School of Health Sciences, Acibadem Mehmet Ali Aydinlar University, Istanbul, Türkiye Izmir International Biomedicine and Genome Institute, Dokuz Eylul University, Izmir, Türkiye Rare and Undiagnosed Disease Platform, IBG – Izmir Biomedicine and Genome Center, Izmir, Türkiye Rare Diseases and Orphan Drugs Application and Research Center (ACURARE), Acibadem Mehmet Ali Aydinlar University, Istanbul, Türkiye Department of Medical Genetics, School of Medicine, Acibadem Mehmet Ali Aydinlar University, Istanbul, Türkiye Department of Child Health and Diseases, School of Medicine, Acibadem Mehmet Ali Aydinlar University, Istanbul, Türkiye Department of Translational Medicine, Graduate School of Health Sciences, Acibadem Mehmet Ali Aydinlar University, Istanbul, Türkiye Department of Medical Biotechnology, Graduate School of Health Sciences, Acibadem Mehmet Ali Aydinlar University, Istanbul, Türkiye Spastic Children’s Foundation of Turkey, Istanbul, Türkiye Department of Pediatric Neurology, School of Medicine, Katip Celebi University, Izmir, Türkiye Pediatric Neurology, Umraniye Training and Research Hospital, University of Health Sciences, İstanbul, Türkiye Department of Pediatric Neurology, School of Medicine, Kocaeli University, Kocaeli, Türkiye Department of Pediatric Neurology, School of Medicine, Bahcesehir University, Istanbul, Türkiye Department of Medical Genetics, Division of Cancer Genetics, Umraniye Training and Research Hospital, Istanbul, Türkiye Department of Neurology, Krankenhaus Nordwest, Frankfurt, Germany Department of Medical Biology, School of Medicine, Acibadem Mehmet Ali Aydinlar University, Istanbul, Türkiye Departments of Neurosurgery and Genetics, Yale Center for Genome Analysis, Yale School of Medicine, New Haven, CT, United States Correspondence: Ugur Ozbek, ugur.ozbek@ibg.edu.tr; Kaya Bilguvar, kaya.bilguvar@acibadem.edu.tr

These authors have contributed equally to this work

 23 02 2026 2026 17 1742186 08 11 2025 13 01 2026 15 01 2026 Copyright © 2026 Yigit, Akgun-Dogan, Ozkeserli, Bayram Akcapınar, Ayta, Gencpinar, Maras Genc, Kutlubay, Kara, Sozen, Agaoglu, Ozdemir, Bilguvar and Ozbek. 2026 Yigit, Akgun-Dogan, Ozkeserli, Bayram Akcapınar, Ayta, Gencpinar, Maras Genc, Kutlubay, Kara, Sozen, Agaoglu, Ozdemir, Bilguvar and Ozbek https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Introduction

Cerebral Palsy (CP) is characterized by permanent, non-degenerative motor function deficits with increasing evidence of genetic contributions. Although prenatal and perinatal risk factors are well recognized, the underlying etiopathology remains incompletely understood. This study aimed to improve diagnostic accuracy and elucidate the genetic architecture of CP and CP-like phenotypes through systematic genomic analyses.

 Methods

Patients with clinically confirmed CP or CP-like presentations were recruited, and biological samples were stored in the ACU-Biobank. Whole-exome and whole-genome sequencing data were analyzed using a validated in-house pipeline incorporating comprehensive variant filtering, prioritization, and re-phenotyping.

 Results

Pathogenic or likely pathogenic variants were identified in 36.4% (24/66) of patients, while variants of uncertain significance (VUS) were detected in 25.8% (17/66). Identified variants involved genes such as SPAST, KIF1A, PLA2G6, CTNNB1, L1CAM, and SYNGAP1. These results demonstrate a substantial contribution of rare monogenic variants to CP and CP-like phenotypes, reflecting extensive genetic heterogeneity.

 Discussion

Our findings support the increasing evidence that genetic factors contribute significantly to CP etiology and emphasize the importance of integrating genomic testing into clinical evaluation. The systematic use of exome and genome sequencing improves diagnostic yield and enables genotype-informed classification, aiding targeted management and genetic counseling for affected individuals.

 Cerebral Palsy CP genetics genomic neurodevelopmental disorders next-generation sequencing The author(s) declared that financial support was received for this work and/or its publication. This study was primarily funded by the Scientific and Technological Research Council of Turkey (TÜBİTAK) under the 1001 Programme (Project no. 221S889, Deep-CP: Development of a Deep Learning Variant Pathogenicity Prediction Tool using AlphaFold for Cerebral Palsy). section-at-acceptance Neurogenetics Introduction

Cerebral Palsy (CP) refers to a broad and diverse group of non-progressive motor disorders characterized by difficulties in movement and posture. Rather than being a single, well-defined condition, CP is a clinical umbrella term for various disorders that share similar symptoms but can stem from different underlying causes (1, 2). These disorders arise from disruptions in brain development during the fetal period or within the first 2 years of life (3). They are often accompanied by comorbidities, such as intellectual disability (ID), epilepsy, visual and/or hearing impairments, and autism spectrum disorder (ASD) (4, 5).

CP is a leading cause of motor disabilities in children, with an estimated prevalence of 1.6 per 1,000 live births in high-income countries (HICs) (6). However, in low- and middle-income countries (LMICs), the numbers are noticeably higher, with rates reaching up to 3.4 per 1,000 live births. This difference highlights the challenges posed by disparities in healthcare access, the quality of perinatal care, and the broader social and economic conditions that shape maternal and infant health outcomes (6).

CP has been associated with a wide range of environmental risk factors occurring during the prenatal, perinatal, and postnatal periods. Nevertheless, the precise pathogenic mechanisms remain incompletely understood, and no single factor has been conclusively identified as a definitive cause (7). In general, the etiology of CP is multifactorial, encompassing environmental, genetic, and epigenetic contributors. Environmental risks include multiple gestations, prematurity, placental abnormalities, intrauterine growth restriction, and perinatal complications such as hypoxic–ischemic injury or infections (8, 9).

In recent years, advances in genomic technologies have highlighted the significant contribution of genetic factors to CP (10). Current research suggests that genetic variants may account for at least one-third of all CP cases (11, 12). The genetic architecture of CP is notably heterogeneous, with new genes continuously being discovered and linked to the condition. This expanding knowledge base emphasizes the critical importance of periodic reanalysis of genomic data. Recently, several studies have increasingly utilized family-based exome sequencing (ES) and genome sequencing (GS) as first-line diagnostic tools to investigate the complex genomic landscape of CP (10, 13). Compared with conventional approaches such as targeted gene panels or chromosomal microarray analysis (CMA), ES and GS provide a more comprehensive genetic assessment, resulting in substantially improved diagnostic yields (14–16). Due to their cost-effectiveness and robust performance, targeted gene panels and ES are currently preferred over GS in many clinical settings. Importantly, ES can generate data on genes not yet associated with disease, which increases the diagnostic potential over time through reanalysis. Indeed, reanalysis of ES data has been shown to uncover previously undiagnosed cases, with studies reporting diagnostic yield increases of 10–25% in patients where initial analyses failed to identify a genetic cause (17–20).

In this study, we aimed to elucidate the molecular etiology of CP cases without identified cause in a Turkish cohort by combining re-phenotyping and reanalysis of genomic data with advanced next-generation sequencing (NGS) approaches, including clinical exome sequencing and short-read whole-genome sequencing. State-of-the-art tools for variant interpretation and prioritization were employed to maximize diagnostic yield.

 Materials and methods Ethics statement

This study was approved by the Acıbadem Mehmet Ali Aydinlar University, and Acıbadem Healthcare Institutions Medical Research Ethics Committee (ATADEK) under decision number 2022–04/18, Spastic Children’s Foundation of Turkey (aka Cerebral Palsy Turkey) academic board, with permission number TSCV.200.023.080 in compliance with the principles of the Declaration of Helsinki (21). Informed consent was obtained from all participants or their legal guardians prior to their inclusion in the study.

 Study cohort

Patients with a preliminary clinical diagnosis of CP were recruited from the Turkey Spastic Children Foundation Clinics, Umraniye Research and Training Hospital Pediatric Genetics and Pediatric Neurology Units, and the Acıbadem Mehmet Ali Aydinlar University, ACURARE Undiagnosed Disease Program—Patient Experience Center between 2018 and 2021. Medical records of patients were retrospectively reviewed.

The study cohort was defined to reduce heterogeneity and potential confounding factors. Patients with a well-defined non-genetic etiology of Cerebral Palsy (CP) were excluded, including those with documented brain injuries such as severe hypoxic–ischemic events (e.g., cord rupture), intraventricular hemorrhage from extreme prematurity, or other acquired perinatal insults. Additionally, individuals with very premature birth (<30 weeks of gestation), severe perinatal asphyxia, known causes like clotting disorders, severe perinatal infections, central nervous system tumors, trauma, or previously confirmed genetic diagnoses were excluded. Patients presenting with prenatal, perinatal, or postnatal risk factors were included only when no definitive non-genetic etiology could be established, as the presence of such risk factors does not preclude an underlying genetic cause. The inclusion criteria consisted of patients aged 0–18 years with a preliminary clinical diagnosis of CP who had undergone clinical exome sequencing, whole-exome sequencing, or whole-genome sequencing without identifying a genetic etiology.

A total of 66 unsolved CP cases were included in the study cohort. Raw data files in Fast Adaptive Shrinkage Thresholding Algorithm and Quality format (FASTQ) obtained from previous clinical exome sequencing (CES), trio whole exome sequencing (WES) and trio whole genome sequencing (WGS) experiments were subsequently obtained. All patients’ medical records, and when needed, in-person assessments of the patients, were assessed by a single pediatric geneticist for deep phenotyping, with the resulting HPO terms utilized in the reanalysis process. Collected data were stored in the Acıbadem Mehmet Ali Aydinlar University, Biobank Unit alongside clinical and demographic information with the informed consent of the patients/ families in accordance with strict patient privacy procedures of the Biobank.

 Genome analysis

FASTQ files of CES, WES and WGS data were processed with an in-house genome analysis pipeline, GENNEXT,1 platform (22). Sequence reads were aligned to the reference genome (GRCh38) independently at each site using SNAP Aligner (23), followed by post-alignment processing with elPrep (24), encompassing duplicate marking, indel realignment, and recalibration of base quality scores. Single-nucleotide variants (SNV) and small indels were called using GATK HaplotypeCaller (25), and DeepVariant Caller (26), annotated using the GENNEXT annotation tool.

 Variant filtering and prioritization

We prioritized variants that met the following criteria: (1) rarity, with a minor allele frequency of 0.0005 (0.005 for recessive inheritance) in population databases such as gnomAD (27, 28) and Turkish Variome (29); (2) variants occurring within coding regions and canonical splice sites; including missense, nonsense, frameshift, start-loss, stop-loss, and canonical splice-site variants, as well as variants predicted to alter splicing; (3) prediction as truncating (including frameshift, stop-gains, and stop-losses), canonical splice sites, missense, or inframe indels by multiple in silico prediction tools, including CADD (30), Polyphen-2 (31), MetaSVM (32), SIFT (33), and SpliceAI (34); (4) integration of curated clinical databases, including OMIM and ClinVar, to assess known gene–disease associations, as identified through comprehensive literature and Genomic England PanelApp;2 and (5) final variant classification performed according to the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP) (35) guidelines, integrating population data, computational evidence, functional consequence, and segregation information when available. The stepwise prioritization workflow is summarized in Supplementary Figure 1.

 Functional enrichment and protein network analysis

To evaluate the functional coherence of the genes prioritized through variant interpretation in the cohort, protein–protein interaction (PPI) network construction and functional enrichment analyses were performed using the STRING web tool (v12.0).3 The analyses were conducted using genes harboring pathogenic/likely pathogenic (P/LP) variants together with clinically relevant variants of uncertain significance (VUS), in order to capture shared functional relationships across the full set of candidate disease-associated genes. PPI enrichment p-values were calculated using a hypergeometric test, comparing the observed number of interactions to that expected by chance for a gene set of similar size.

For functional enrichment, StringDB identifies over-represented Gene Ontology (GO) terms, KEGG pathways,4 and disease associations, applying Benjamini–Hochberg correction to control the false discovery rate (FDR). The same gene set (P/LP + clinically relevant VUS) was used for GO enrichment analysis. Terms with corrected FDR < 0.05 were considered significantly enriched. In addition, enrichment strength scores, defined as the log₁₀ ratio of observed to expected gene counts, were used to quantify effect size, and terms with a strength greater than 0.5 were reported to highlight biologically meaningful associations. All statistical analyses were performed as implemented in the STRING web platform, using its default settings unless otherwise specified.

 Results CP cohort characteristics

A total of 66 patients were included in the study. The cohort comprised 32 females and 34 males (Figure 1A). The mean age of the cohort was 6.0 years, ranging from 3 months to 18 years. Consanguinity was identified in approximately 44% (29/66) of families (Figure 1B).

Clinical and demographic characteristics of the cohort. (A) Distribution of sex among the patients (male: 52%, female: 48%). (B) Parental consanguinity rate, with 43.9% of patients born to consanguineous parents and 56.1% to non-consanguineous parents. (C) Frequency of comorbidities observed in the cohort. The most common comorbidity was developmental delay (n = 60), followed by congenital abnormalities (n = 40), epilepsy (n = 38), and visual and/or hearing impairment (n = 15). Less common comorbidities included ataxia (n = 6), autism spectrum disorders (n = 4), and myopathy (n = 3). (D) Identified risk factors (n = 34) were classified according to the timing of occurrence as antenatal (52.9%), natal (11.8%), or postnatal (35.3%). Classification was based on whether the risk factor occurred before delivery (antenatal), during labor and delivery (natal), or after birth (postnatal). Percentages represent event-based frequencies, as more than one risk factor could be present in the same patient.

Four-part data visualization. A) Sex distribution: 52% male, 48% female. B) Consanguinity rate: 43.9% consanguineous, 56.1% non-consanguineous. C) Comorbidities: Developmental delay, congenital abnormalities, and epilepsy affect 60, 40, and 38 patients respectively, while visual/hearing impairment, ataxia, autism spectrum disorders, and myopathy affect 15, 6, 4, and 3 patients. D) Risk factors: Antenatal 52.9%, postnatal 35.3%, natal 11.8%.

All patients presented with at least one comorbidity. Developmental delay was the most frequently reported comorbidity, observed in more than half of the cohort, followed by congenital anomalies and epilepsy. The reported comorbidities and their distribution are illustrated in Figure 1C. Detailed clinical findings and corresponding HPO terms for each individual are provided in Table 1; Supplementary Table 1 summarizes the demographic and clinical characteristics of the cohort.

Clinical details of Turkish CP cohort.

Patient ID Sex Age at last examination Cons. Phenotypic features
CP_P1.1 F 5,0 Yes Facial dysmorphism, global developmental delay, ataxia, cerebellar vermis hypoplasia, periventricular white matter hyperintensities and ectopic posterior pituitary
CP_P2.1 M 9,0 No Gastroesophageal reflux, global developmental delay, seizures, dysphagia, hypoplasia of the corpus callosum, ventriculomegaly, frontotemporal cerebral atrophy, and calcification of the globus pallidus
CP_P3.1 F 11,0 NA Scoliosis, global developmental delay, hypotonia, and intellectual disability
CP_P4.1 M 11,2 Yes Global developmental delay, motor delay and intellectual disability
CP_P5.1 F 4,1 No Global developmental delay, hypotonia, ataxia, cerebral atrophy, and drooling
CP_P6.1 F 0,6 NA Facial dysmorphism, nuchal translucency, biparietal narrowing, and neonatal hypotonia
CP_P7.1 F 9,0 Yes Facial dysmorphism, lower limb spasticity, myopathy, muscular dystrophy, distal prominence of the metatarsals, global developmental delay, hypotonia, seizures, dysphagia, dysplasia of the corpus callosum, and elevated serum creatine phosphokinase
CP_P8.1 M 0,6 NA Distal arthrogryposis
CP_P9.1 M 9,0 NA Facial dysmorphism, neonatal asphyxia, scoliosis, spastic diplegia, pes planus, cryptorchidism, hyperreflexia and feeding difficulties
CP_P10.1 F 0,3 Yes The high palate, neonatal respiratory distress, talipes equinovarus, facial hypotonia, polyneuropathy, and periventricular white matter hyperintensities
CP_P11.1 M 12,0 NA Microcephaly, and intellectual disability
CP_P12.1 M 2,0 NA Coarse facial features, global developmental delay, seizures, widespread papulopustular rash on the anterior and posterior surfaces of the chest
CP_P13.1 F 7,3 Yes Microcephaly, global developmental delay, motor delay, seizures, spasticity, absent speech, bruxism, hyperacusis, dysmorphic facial features, difficulty walking, limited knee, elbow flexion, and periventricular white matter hyperintensities
CP_P14.1 F 1,2 Yes Short neck, hypertelorism, scoliosis, global developmental delay, hypotonia, seizures and spasticity
CP_P15.1 M 2,9 No Global developmental delay, absent speech, seizures, spasticity, hydrocephalus, intellectual disability, motor delay, cerebral atrophy, bruxism, and hypoplasia of the corpus callosum
CP_P16.1 M 6,0 No Facial dysmorphism, retrognathia, hypertelorism, talipes equinovarus, pes cavus, hypospadias, cryptorchidism, global developmental delay, delayed speech and language development, and agenesis of the corpus callosum
CP_P17.1 F 12,8 No Bilateral hearing impairment, poor head control, global developmental delay, delayed speech and language development, learning difficulties, motor delay, increased deep tendon reflexes, inguinal hernia, thrombocytopenia, proteinuria and hematuria
CP_P18.1 M 5,3 Yes Strabismus, global developmental delay, seizures, motor delay, neurodevelopmental delay, deficiency in independent sitting, chronic infection and continuous fever
CP_P19.1 F 8,8 Yes Microcephaly, intellectual disability, long palpebral fissures, macrodontia, global developmental delay, delayed speech and language development, and seizures
CP_P20.1 F 2,4 No Pes planus, global developmental delay, motor delay, epileptic encephalopathy (under follow-up with a diagnosis of CP)
CP_P21.1 M 5,6 No Retinal detachment, and hypotonia
CP_P22.1 F 7,1 Yes Microcephaly, hypertonia, nephrotic syndrome, global developmental delay, seizures, spasticity, and elevated serum creatine phosphokinase
CP_P23.1 F 1,1 No Global developmental delay, hypotonia, seizures and muscle spasms
CP_P24.1 M 1,7 Yes Microcephaly, facial dysmorphism, neonatal hypotonia, hypertonia, seizures, agenesis of the corpus callosum, hyperreflexia, dilation of lateral ventricles and neurodevelopmental delay
CP_P25.1 F NA Yes Microcephaly, pectus excavatum, seizures, motor delay and frontal cortical atrophy
CP_P26.1 F 0,4 No Retrognathia, respiratory distress, flexion contracture, limited knee extension, wide intermammillary distance, arthrogryposis multiplex congenita, and poor suck
CP_P27.1 M 8,0 No Retinitis pigmentosa inversa, hearing impairment, respiratory distress, arthritis, global developmental delay, delayed speech and language development, seizures, and ataxic gait
CP_P28.1 M 11,8 Yes Global developmental delay, autism, absent speech, seizures, motor delay, polyneuropathy, and neurodevelopmental delay
CP_P29.1 M 0,5 Yes Microcephaly, facial dysmorphism, visual impairment, tapered finger, global developmental delay, seizures, dystonia, motor delay, EEG abnormalities, diffuse cerebral atrophy, striatal T2 hyperintensity, dystonic gait, epileptic encephalopathy, and congenital hypothyroidism
CP_P30.1 F 5,4 Yes Poor suck, scoliosis, torticollis, neonatal hypotonia, kyphosis, hip dislocation, myopathy, muscular dystrophy, and joint hypermobility
CP_P31.1 F 10,7 Yes High narrow palate, strabismus, hypertelorism, difficulty walking, clinodactyly, global developmental delay, delayed speech and language development
CP_P32.1 F 8,7 Yes Microcephaly, global developmental delay, neurodevelopmental delay
CP_P32.2 M 3,6 Yes Microcephaly, global developmental delay, neurodevelopmental delay
CP_P33.1 M 0,3 Yes Neonatal asphyxia, hypotonia and dysphagia
CP_P34.1 M 13,2 Yes Facial asymmetry, seizures, homocystinuria, abnormalities of glutamine and homocysteine metabolism
CP_P35.1 M 9,0 No Microcephaly, tracheomalacia, bronchomalacia, spastic tetraplegia, cryptorchidism, seizures, cerebellar atrophy, infantile spasms, and feeding difficulties
CP_P36.1 F 7,7 No Nystagmus, strabismus, difficulty walking, global developmental delay, delayed speech and language development, seizures and agenesis of the corpus callosum
CP_P37.1 F 1,8 No Retrognathia, facial dysmorphism, visual impairment, abnormal electroretinogram, neonatal hypotonia, seizures, motor delay, cerebral atrophy, increased deep tendon reflexes, global developmental delay, milk allergy and feeding difficulties
CP_P38.1 F 7,0 Yes Microcephaly, intellectual disability, global developmental delay, seizures and autism
CP_P39.1 M 7,0 No Facial dysmorphism, strabismus, prominent fingertip pads, global developmental delay, motor delay, polyhydramnios and prematurity
CP_P40.1 M 1,7 Yes Brachycephaly, atrial septal defect, ventricular septal defect, autism, hypotonia, motor delay, self-injurious behavior, and inappropriate laughter
CP_P41.1 F 1,9 Yes Global developmental delay, hypotonia, seizures, cerebellar vermis atrophy, dilation of lateral ventricles, aggressive behavior, inferior cerebellar vermis hypoplasia, abnormality of glycolysis and abnormality of creatine metabolism
CP_P42.1 F 2,9 No Global developmental delay, delayed speech and language development, seizures, and EEG abnormality
CP_P43.1 M 5,1 No Neonatal sepsis, microcephaly, seizures, delayed speech and language development, global developmental delay
CP_P44.1 M 4,9 No Cerebral Palsy
CP_P45.1 M 3,1 No Growth retardation, strabismus, scoliosis, pes cavus, abnormality of toe proximal phalanx, global developmental delay, spasticity and cerebellar vermis atrophy
CP_P46.1 M 8,4 Yes Hypertonia, lower limb spasticity, global developmental delay, stereotypes, and EEG abnormality
CP_P47.1 F 1,8 No Macrocephaly, facial dysmorphism, hypotonia, hydrocephalus, convulsive status epilepticus, dilation of lateral ventricles, and periventricular leukomalacia
CP_P48.1 M 10,7 No Mild global developmental delay, seizures, cerebellar atrophy, agenesis of corpus callosum, reduced tendon reflexes, neonatal hypotonia, poor suck, cerebral atrophy, gait ataxia, hypoplasia of the corpus callosum, febrile seizures, dilation of lateral ventricles
CP_P49.1 M 2,9 No Mild microcephaly, impaired mastication, laryngomalacia, pes planus, infantile axial hypotonia, delayed speech and language development, motor delay, reduced tendon reflexes, delayed myelination, and hyperintensity of cerebral white matter on MRI
CP_P50.1 M 7,8 Yes Hypodontia, strabismus, nystagmus, difficulty walking, global developmental delay, seizures, motor delay and EEG abnormality
CP_P51.1 M 2,8 No Delayed gross motor development, increased nuchal translucency, delayed speech and language development, nystagmus, astigmatism, multiple tooth decay, relative macrocephaly, polyneuropathy, hypotonia and pectus carinatum
CP_P52.1 F 9,0 No Global developmental delay, seizures, ataxia, severe motor delay, absent speech, Intellectual disability, pontocerebellar hypoplasia, facial dysmorphism, vomiting
CP_P53.1 F 3,0 No Global developmental delay, poor suck, Cerebral Palsy, axial hypotonia, hypertonia, inability to walk, absent speech, stereotypical hand wringing
CP_P54.1 F 3,0 No Microcephaly, hypertonia, poor head control, global developmental delay, cerebellar atrophy, hypoplasia of the corpus callosum, dilation of lateral ventricles, generalized hypotonia, seizures, dystonia
CP_P55.1 M 7,0 No Global developmental delay, seizures, hand tremor, Inability to walk, absent speech, hydrocephalus
CP_P55.2 F 2,0 No Global developmental delay, seizures, EEG abnormality
CP_P56.1 M 6,0 No Motor delay, spasticity, delayed speech and language development, delayed gross motor development, difficulty walking, hyperactive deep tendon reflexes, broad hallux, pes cavus
CP_P57.1 F 11,0 No Visual impairment, upslanted palpebral fissures, seizures, cerebellar atrophy, generalized hypotonia, lower limb spasticity, delayed gross motor development, difficulty walking, knee flexion contracture, cerebellar vermis atrophy, pes valgus
CP_P58.1 F 6,0 Yes Neonatal respiratory distress, brachycephaly, retrognathia, narrow forehead, proptosis, absent speech, seizures, spasticity, hydrocephalus, partial agenesis of corpus callosum, EEG abnormality, inability to walk, colpocephaly
CP_P59.1 M 0,7 Yes Global developmental delay, generalized hypotonia, round face, poor head control, spasticity
CP_P60.1 F 3,0 Yes Facial dysmorphism, microcephaly, hypertonia, rocker bottom root, global developmental delay, motor delay, generalized tonic–clonic seizures, ventriculomegaly, delayed myelination, generalized hypotonia
CP_P61.1 M 9,0 No Seizures, intellectual disability, spastic paraplegia, spasticity, global developmental delay, delayed ability to walk, difficulty walking,
CP_P62.1 M 18,0 Yes Stereotypy, seizures, global developmental delay, spastic tetraparesis, tremor, absent speech, EEG abnormality, difficulty walking, hyperactive deep tendon reflexes, neurodevelopmental delay, delayed ability to walk
CP_P63.1 F 18,0 Yes Global developmental delay, mild intellectual disability, delayed speech and language development, spasticity, difficulty walking, pes cavus, joint laxity
CP_P64.1 M 10,0 No Global developmental delay, seizures, intellectual disability, motor regression, developmental regression, absent speech, inability to walk, EEG abnormality

Most patients had no documented history of risk factors. However, 37.9% (25 of 66) of the cohort had at least one identified risk factor associated with CP, including prematurity, small for gestational age, postnatal respiratory distress, polyhydramnios, maternal vaginal bleeding, breech delivery, neonatal sepsis, or intrauterine growth restriction (Figure 1D, summarizing prenatal, natal, and postnatal risk factor events identified in the cohort).

 Genome analysis

Genomic analyses were performed in a cohort of 66 unsolved CP and CP-like cases using two complementary strategies. For 57 individuals, existing clinical exome sequencing (CES) or whole-exome sequencing (WES) data was conducted, depending on data availability and quality. In 13 cases, newly generated trio-based whole-exome sequencing (WES) or whole-genome sequencing (WGS) data were analyzed, including unaffected parents and, when available, affected or unaffected siblings (Figure 2).

Overview of genomic analysis strategies and diagnostic outcomes in the study cohort. Genomic analyses were performed in a cohort of 66 patients with CP and CP-like phenotypes identified through hospital records from multiple centers. Two complementary approaches were applied: reanalysis of previously generated exome sequencing (CES or WES) data (n = 57) and analysis of newly generated trio-based WES or WGS data (n = 13). P/LP variants fulfilling ACMG/AMP criteria were considered to provide a definitive molecular diagnosis, while VUSs were reported separately as clinically relevant candidate findings. Cases without a definitive molecular diagnosis were classified as genetically unresolved.

Flowchart illustrating the analysis of patients with CP identified from hospital records. Sixty-six patients were divided into two groups: fifty-seven underwent reanalysis of previous tests, resulting in 29.8% solved P/LP and 24.5% clinically relevant VUS. Nine patients with no prior sequencing received WES/WGS data analysis, yielding 53.8% solved P/LP and 23.1% clinically relevant VUS. Overall, 37.8% of cases had negative results.

Using these complementary approaches, a definitive molecular diagnosis based on the identification of pathogenic or likely pathogenic (P/LP) variants was established in 24 of 66 patients (36.4%). Table 2 summarizes the pathogenic and likely pathogenic variants found in individuals with a definitive genetic diagnosis, providing detailed annotations of variant types and their ACMG classifications.

Pathogenic and likely pathogenic variants identified in the cohort and clinical and genetic information of patients harboring these variants.

Patient ID Gene Variant GnomAD v4 allele Freq. variant type DISORDER (MIM) CP risk factors Zyg (segregation) ACMG classification
CP_P17.1 ASXL1 SETBP1 NM_015338.6: c.2077C > T (p. Arg693Ter)NM_015559.3: c.2612 T > C (p. Ile871Thr) 0.000016.196e-7 Stop gainMissense Bohring-Opitz syndrome (MIM #605039) and Intellectual developmental disorder, autosomal dominant 29 (MIM #616078)Schinzel-Giedion midface retraction syndrome (MIM #269150) Low birth weight Het (NA)Het (NA) P (PVS1, PS4, PM2, PP5)P (PM2, PS4, PM5, PM1, PP3, PP5)
CP_P39.1 PPM1D NM_003620.4:c.1262C > A(p. Ser421Ter) 6.300e-7 Stop gain jansen-de Vries syndrome (MIM #617450) Polyhydramnios, vaginal bleeding, prematurity and low birth weight Het (NA) P (PVS1, PS4, PM2, PP5)
CP_P8.1 COL6A1 NM_001848.3: c.1693C > T (p. Arg565Ter) 0.000001 Stop gain Ulrich congenital muscular dystrophy 1A (MIM #254090) – Hom (NA) P (PVS1, PM2, PS4, PP5)
CP_P21.1 CTNNB1 NM_001904.4: c.1923dup (p. Glu642ArgfsTer6) – Frameshift Exudative vitreoretinopathy 7 (MIM #617572) Neurodevelopmental disorder with spastic diplegia and visual defects (MIM #615075) – Het (NA) P (PVS1, PM2, PS4, PP5)
CP_P30.1 COL6A1 NM_001848.3: c.811C > T (p. Arg271Ter) 0.000006 Stop gain Ulrich congenital muscular dystrophy 1A (MIM #254090) Breech presentation at birth Hom (NA) P (PVS1, PM2, PP5)
CP_P46.1 CACNA1E NM_001205293.3:c.5365-3_5,365-2del – Splice site Developmental and epileptic encephalopathy 69 (MIM #618285) – Het (de novo) P (PVS1, PM2, PM6)
CP_P59.1 NGLY1 NM_018297.4:c.1533_1536del (p. Asn511Lysfs*51) 0.00001 Frameshift Congenital disorder of deglycosylation 1 (MIM #615273) – Hom (maternal paternal) P (PVS1, PM2, PM3, PP5)
CP_P61.1 ATP1A3 NM_152296.5:c.2324C > T (p. Pro775Leu) – Missense Developmental and epileptic encephalopathy 99 (MIM #619606) CAPOS syndrome (MIM #601338) – Het (NA) P (PS2, PM2, PM5, PM1, PP3, PP2, PP5)
CP_P14.1 WWOX NM_016373.4: c.716 T > G (p. Leu239Arg) – Missense Developmental and epileptic encephalopathy 28 (MIM #616211) Spinocerebellar ataxia, autosomal recessive 12 (MIM #614322) Intrauterine growth retardation Hom (maternal paternal) P (PP3, PM3, PM2, PP5)
CP_P13.1 SYNGAP1 NM_006772.3: c.851 T > G (p. Leu284Arg) – missense Intellectual developmental disorder, autosomal dominant 5 (MIM #612621) Polyhydramnios and prolonged labor Het (de novo) P (PP3, PM2, PM5, PP2, PS2)
CP_P2.1 PLA2G6 NM_003560.4: c.1772G > A (p. Arg591Gln) 0.000008 Missense Infantile neuroaxonal dystrophy (MIM #256600) Neurodegeneration with brain iron accumulation 2B (MIM #610217) – Hom (NA) P (PM3, PM2, PM5, PM1, PP3, PP2, PP5)
CP_P62.1 ST3GAL5 NM_003896.4:c.601G > A (p. Gly201Arg) 0.000009 Missense Salt and pepper developmental regression syndrome (MIM #609056) – Hom (maternal paternal) P (PM2, PP3, PP5)
CP_P57.1 KIF1A NM_001244008.2:c.946C > T (p. Arg316Trp) – Missense NESCAV syndrome (MIM #614255) Spastic paraplegia 30, autosomal dominant (MIM #610357) Low birth weight Het (de novo) P (PM2, PM5, PM1, PS2, PP3, PP2, PP5)
CP_P64.1 STXBP1 NM_001032221.6:c.578G > C (p. Gly193Ala) – Missense Developmental and epileptic encephalopathy 4 (MIM #612164) – Het (de novo) P (PM2, PM5, PM1, PP3, PP2, PP5, PS2)
CP_P12.1 ARID1B NM_001374828.1:c.1391_1400del (p. Gly464fs) – Frameshift Coffin-Siris syndrome 1 (MIM #135900) Neonatal seizures Het (NA) LP (PVS1, PM2)
CP_P33.1 COL1A2 NM_000089.4:c.3275dup (p. Gly1093TrpfsTer17) 0.000001 Frameshift Combined osteogenesis imperfecta and Ehlers-Danlos syndrome 2 (MIM #619120) Ehlers-Danlos syndrome, arthrochalasia type,2 (MIM #617821) Neonatal respiratory distress Het (NA) LP (PVS1, PM2)
CP_P40.1 LINS1 NM_001040616.3:c.1870del (p. Gln624LysfsTer4) – Frameshift Intellectual developmental disorder, autosomal recessive 27 (MIM #614340) – Hom (NA) LP (PVS1, PM2)
CP_P31.1 SAMD9 NM_017654.4:c.2159del (p. Asn720ThrfsTer35) 0.000009 Frameshift MIRAGE syndrome (MIM #617053) Vaginal bleeding, and low birth weight Het (NA) LP (PVS1, PM2, PP5)
CP_P48.1 L1CAM NM_001278116.2:c.414 > A (p. Trp138Ter) – Stop gain Corpus callosum, partial agenesis of (MIM #304100) Hydrocephalus, congenital, X linked (MIM #307000) MASA syndrome (MIM #303350) – Hem (de novo) LP (PVS1, PM2, PM6)
CP_P47.1 PTEN NM_000314.8:c.275A > T (p. Asp92Val) – Missense Cowden syndrome 1 (MIM #158350) – Het (NA) LP (PP3, PM2, PM5, PM1, PP2, PP5, PS3)
CP_P42.1 PIK3R2 NM_005027.4: c.1117G > A (p. Gly373Arg) – Missense Megalencephaly-polymicrogyria-polydactyly-hydrocephalus syndrome 1 (MIM #603387) – Het (NA) LP (PM2, PS4, PM1, PP5)
CP_P51.1 MFN2 NM_014874.4: c.743 T > C (p. Leu248Pro) – Missense Charcot–Marie–Tooth disease, axonal, type 2A2A (MIM #609260) Hereditary motor and sensory neuropathy VIA (MIM #601152) – Het (de novo) LP (PM1, PP2, PM2, PM5, PP3, PM6)
CP_P53.1 GRIN2B NM_000834.5:c.2021C > T (p. Pro674Leu) – Missense Developmental and epileptic encephalopathy 27 (MIM #616139) Intellectual developmental disorder, autosomal dominant 6, with or without seizures (MIM #613970) – Het (de novo) LP (PM1, PM2, PP2, PM6)
CP_P56.1 SPAST NM_014946.4:c.1496G > T (p. Arg499Leu) 6.222e-7 Missense Spastic paraplegia 4, autosomal dominant (MIM #182601) – Het (maternal) LP (PM1, PM2, PM5, PP2, PP3, PP5)

Het, Heterozygous; Hem, hemizygous; Hom, Homozygous; P, Pathogenic; LP, Likely pathogenic; NA, Not applicable.

In an additional 17 patients (25.8%), variants of uncertain significance (VUS) were identified. Although these variants did not meet ACMG criteria for pathogenic or likely pathogenic classification, they were considered clinically relevant candidate variants due to strong genotype–phenotype concordance, segregation data when available, and consistency with previously reported disease mechanisms. These cases were therefore not regarded as definitive molecular diagnoses but were reported separately as putative genetic findings (Table 3).

Variants of uncertain significance (VUS) identified in the cohort and clinical and genetic information of patients harboring these variants.

Patient ID Gene Variant GnomAD v4 allele Freq. Variant type DISORDER (MIM) CP risk factors Zyg (segregation) ACMG classification
CP_P15.1 CLIC2 NM_001289.6: c.103C > T (p. Arg35Cys) 0.00001 Missense Intellectual developmental disorder, X-linked syndromic 32 (MIM #300886) Hem (NA) VUS (PM2)
CP_P58.1 KCNT1 NM_020822.3:c.3571G > A (p. Glu1191Lys) 0.000004 Missense Developmental and epileptic encephalopathy 14 (MIM #614959) – Hom (maternal paternal) VUS (PM2)
CP_P3.1 GATM NM_001482.3: c.1081A > G (p. Asn361Asp) – Missense Cerebral creatine deficiency syndrome 3 (MIM #612718) – Hom (NA) VUS (PM2, PP3)
CP_P4.1 GRM1 NM_001278064.2: c.1124G > A (p. Arg375Gln) 6.196e-7 Missense Spinocerebellar ataxia, autosomal recessive 13 (MIM #614831) – Hom (NA) VUS (PM2, PP3)
CP_P28.1 NDUFS3 NM_004551.3:c.252A > T (p. Leu84Phe) 0.00002 Missense Mitochondrial complex I deficiency, nuclear type 8 (MIM #618230) – Hom (NA) VUS (PM2, PP3)
CP_P29.1 CHKA NM_001277.3:c.630G > C (p. Pro210=) 6.196e-7 Splice site Neurodevelopmental disorder with microcephaly, movement abnormalities, and seizures (MIM #620023) – Hom (maternal paternal) VUS (PM2, PP3)
CP_P37.1 ST3GAL5 NM_003896.4: c.722G > C (p. Arg241Thr) – Missense Salt and pepper developmental regression syndrome (MIM #609056) – Hom (maternal paternal) VUS (PM2, PP3)
CP_P50.1 KCNQ2 NM_172107.4:c.1526-3C > A 6.209e-7 Splice site Developmental and epileptic encephalopathy 7 (MIM #613720) Neonatal seizures Hom (maternal paternal) VUS (PM2, PP3)
CP_P1.1 SCN2A NM_001040142.2: c.5726C > T (p. Ala1909Val) 0.000001 Missense Developmental and epileptic encephalopathy 11 (MIM #613721) – Het (NA) VUS (PM2, PP3, PP2)
CP_P35.1 CACNA1G NM_018896.5:c.4987G > C (p. Val1663Leu) 0.000002 Missense Spinocerebellar ataxia 42 (MIM #616795) Spinocerebellar ataxia 42, early-onset, severe, with neurodevelopmental deficits (MIM #618087) – Het (NA) VUS (PM2, PP3, PP2)
CP_P38.1 CACNA1G NM_018896.5: c.1211G > A (p. Arg404Gln) 0.000001 Missense Spinocerebellar ataxia 42 (MIM #616795) Spinocerebellar ataxia 42, early-onset, severe, with neurodevelopmental deficits (MIM #618087) – Het (NA) VUS (PM2, PP3, PP2)
CP_P25.1 KCNT1 NM_020822.3:c.2756C > T (p. Thr919Met) – Missense Developmental and epileptic encephalopathy 14 (MIM #614959) – Het (NA) VUS (PM2, PP3, PM1)
CP_P6.1 SCN8A NM_014191.4: c.3352G > C (p. Asp1118His) – Missense Cognitive impairment with or without cerebellar ataxia (MIM #614306) Developmental and epileptic encephalopathy 13 (MIM #614558) – Het (NA) VUS (PM2, PP2)
CP_P36.1 PACS1 NM_018026.4:c.1775G > A (p. Cys592Tyr) – Missense Schuurs-Hoeijmakers syndrome (MIM #615009) Neonatal seizures Het (NA) VUS (PM2, PP2)
CP_P23.1 ATP1A2 NM_000702.4:c.736A > G (p. Asn246Asp) 0.00001 Missense Alternating hemiplegia of childhood 1 (MIM #104290) Developmental and epileptic encephalopathy 98 (MIM #619605) – Het (NA) VUS (PM2, PP2, PP3)
CP_P18.1 PNPT1 NM_033109.5:c.2212C > T(p. Arg738Cys) 0.00001 Missense Spinocerebellar ataxia 25 (MIM #608703) Low birth weight Het (NA) VUS (PM2, PM5)
CP_P7.1 ACY1 NM_000666.3:c.1057C > T (p. Arg353Cys) 0.003 Missense Aminoacylase 1 deficiency (MIM #609924) – Hom (NA) VUS (BS1, PP5)

Het, Heterozygous; Hem, hemizygous; Hom, Homozygous; VUS, Variant of Uncertain Significance; NA, Not applicable.

When considered together, pathogenic/likely pathogenic variants and clinically relevant VUS accounted for molecular findings in 41 of 66 patients (62.1%), reflecting the overall genetic landscape of the cohort. However, only cases harboring P/LP variants were classified as having a definitive genetic diagnosis, while cases with VUS remain under ongoing evaluation (Figure 3A).

Genetic diagnosis status and distribution of disorder groups. (A) Genetic diagnostic results in the CP patient group. “Solved (P/LP)” indicates cases with pathogenic or likely pathogenic variants (36.4%), while “Clinically relevant VUSs” includes cases with variants of uncertain significance (25.8%). The remaining 37.9% of cases remain genetically unresolved. (B) Distribution of disease categories among genetically diagnosed cases. The largest group is Neurodevelopmental Disorders (50.0%), followed by Syndromic Disorders (26.2%), Metabolic Disorders (9.5%), Neuromuscular Disorders (9.5%), and Hereditary Spastic Paraplegias (4.8%).

Chart A is a donut chart showing variant unsolved at 37.9%, solved (P/LP) at 36.4%, and clinically relevant VUSs at 25.8%. Chart B is a pie chart detailing disorders: neurodevelopmental at 50.0%, syndromic at 26.2%, metabolic at 9.5%, neuromuscular at 9.5%, and hereditary spastic paraplegias at 4.8%.

In total, 42 variants were identified across these 41 individuals. Missense variants constituted the majority (66.6%, 28/42), followed by truncating variants including frameshift and stop-gain variants (26.2%, 11/42), and canonical splice-site variants (7.1%, 3/42). No pathogenic mitochondrial DNA variants or clinically relevant copy number variants were detected in this subset.

Among the 42 genetic diagnoses in 41 solved cases, autosomal dominant inheritance was most common (59.5%), followed by autosomal recessive (35.7%, all homozygous) and X-linked (4.76%). One patient had dual diagnoses involving distinct phenotypes.

Patients were classified into five phenotype-based groups based on their underlying molecular etiologies. Neurodevelopmental disorders were the most frequent (50.0%), followed by syndromic disorders (26.2%), neuromuscular disorders (9.5%), metabolic disorders (9.5%), and hereditary spastic paraplegias (4.8%) (Figure 3B).

The genes identified as harboring potential disease susceptibility variants in the Turkish patient cohort were analyzed for network and pathway enrichment using the STRING web tool (v12.0). The STRING network analysis of these genes yielded a PPI enrichment p-value of 9.08 × 10−13, indicating that at least a part of the proteins is functionally related and likely to participate in shared biological pathways or processes (Supplementary Figure 2).

Gene Ontology enrichment revealed strong over-representation of neurophysiological processes, including membrane depolarization during action potential (FDR = 1.34 × 10−13) and glutamatergic synaptic transmission (FDR = 2.46 × 10−12) (Supplementary Figure 3; Supplementary Table 3). Molecular function analysis highlighted voltage-gated ion channel activity (FDR = 1.85 × 10−13) and metal ion transmembrane transporter activity (FDR = 5.47 × 10−12) (Supplementary Figure 4; Supplementary Table 4).

Cellular component terms were dominated by neuronal and membrane-associated structures, such as neuron projection (FDR = 2.93 × 10−14), synapse (FDR = 1.07 × 10−13), and voltage-gated sodium channel complex (FDR = 2.85 × 10−12) (Supplementary Figure 5; Supplementary Table 5). KEGG pathway analysis identified enrichment in cAMP signaling (FDR = 3.21 × 10−5), Ras signaling (FDR = 7.62 × 10−5), and thyroid hormone signaling (FDR = 1.34 × 10−4) (Supplementary Table 6; Supplementary Figure 6). Disease enrichment analysis linked the prioritized genes most strongly to epilepsy (FDR = 5.21 × 10−16), autism spectrum disorder (FDR = 8.14 × 10−15), and intellectual disability (FDR = 3.66 × 10−14) (Supplementary Table 7; Supplementary Figure 7).

Taken together, these findings support a mechanistic link between the selected variants and central nervous system dysfunction, particularly processes affecting ion homeostasis, neuronal excitability, and neurodevelopmental pathologies (Figure 4 and all raw data are provided in Supplementary material).

Functional enrichment analysis of prioritized genes in the cohort. Biological process (gene ontology) enriched in analysis conducted in STRING web tool (v12.0). The plot displays the top 16 enriched GO terms ranked by the highest signal scores and lowest FDR values, grouped by semantic similarity (threshold ≥0.8). Signal is a composite measure prioritizing enriched terms based on statistical significance (adjusted p-value) and informativeness. While the strength score (Supplementary material) represents the effect size, calculated as the log₁₀ ratio of observed to expected gene counts. Each bar corresponds to the signal magnitude for an individual GO term, and the bubbles at the end of each bar denote the number of genes in the query set that overlap with genes annotated with that GO term. Statistical significance of each overlap is assessed using a hypergeometric test, and p-values are corrected for multiple testing using the Benjamini–Hochberg method, reported as FDR (false discovery rate). A GO term with a lower signal may still appear higher in the plot if it is semantically grouped with a higher-ranking term. Background shading reflects the FDR of each group.

Bubble chart representing various biological processes against signal values. Processes include "Membrane depolarization during action potential" and others, with signal values ranging from 0.6 to 1.1. Bubble size represents gene count, with larger bubbles indicating more genes. Color gradient signifies FDR values, from deep purple for lower values to orange for higher values. Discussion

Genomic testing is transforming the diagnostic landscape of CP, revealing that many cases previously considered idiopathic have an underlying genetic basis. CP includes a wide range of motor and neurodevelopmental impairments, yet identifying a clear genetic cause remains difficult in many instances. In our study, we highlight the transformative role of genomic analyses in uncovering underlying molecular mechanisms, which pave the way for personalized treatment strategies, informed genetic counseling, and tailored long-term follow-up.

In this study, the use of advanced bioinformatics pipelines combined with deep phenotyping and diverse genomic sequencing approaches yielded a diagnostic rate of approximately 38%, enabling the identification of genetic etiologies in previously undiagnosed CP patients. Reported diagnostic yields in prior genomic studies range from 24 to 58%, influenced by factors such as the inclusion of idiopathic cases, the presence of neurodevelopmental comorbidities, and the use of trio-based sequencing strategies (10, 15, 16, 36).

Our high diagnostic yield is likely due to several factors: careful patient selection based on strict inclusion and exclusion criteria, the use of comprehensive deep phenotyping, and the deliberate exclusion of individuals with CP clearly linked to non-genetic causes (e.g., perinatal asphyxia, extreme prematurity). Additionally, integrating advanced bioinformatics pipelines with thorough segregation analyses enhanced our ability to identify genetic causes that had previously gone undetected in this patient group. CP shares significant phenotypic overlap with other neurodevelopmental disorders, and many genetic neurological conditions can resemble CP because of similar motor and developmental features (37). In our cohort, the variety of underlying genetic causes underscored the diverse molecular mechanisms involved in CP and CP-like presentations. Reverse phenotyping helped refine diagnoses into five main groups (Tables 2, 3), with neurodevelopmental disorders, particularly epileptic encephalopathies [e.g., WWOX (MIM: 605131), ATP1A3 (MIM: 619606)] and spinocerebellar ataxia [e.g., PNPT1 (MIM: 610316)], forming the largest subgroup. Variants in genes such as L1CAM (MIM: 304840), CTNNB1 (MIM: 116806), and SYNGAP1 (MIM: 603384), recently suggested as candidates in CP pathophysiology (36), are involved in neuronal migration, synaptogenesis, and axonal guidance (1, 38). These findings support the idea that CP and related neurodevelopmental disorders may exist on a shared clinical and molecular spectrum, highlighting the importance of genomic testing in clarifying diagnostic boundaries where phenotypic overlap complicates classification.

Although all patients presented with at least one reported comorbidity, these features should not necessarily be interpreted as independent conditions co-occurring with Cerebral Palsy. In many cases, they represent core manifestations of the underlying genetic disorder. For example, epilepsy in patients with developmental and epileptic encephalopathies constitutes a primary disease feature rather than a secondary comorbidity. Similarly, visual impairment in CTNNB1-related (MIM: 116806) disorders and ataxia in CACNA1G-associated (MIM: 604065) conditions are well-established components of their respective phenotypic spectrum.

The second largest group comprised syndromic cases, where dysmorphic features especially in patients with severe neurodevelopmental impairments were often overlooked under an initial CP diagnosis (39). Such oversight can lead to misclassification and missed phenotypic clues essential for accurate genetic evaluation. In our study, incomplete phenotyping and underreporting of HPO terms contributed to missed variants in syndromic genes during initial panel-based analyses (40). Reverse phenotyping later revealed unrecognized dysmorphic features, such as those associated with ARID1B (Coffin–Siris syndrome, MIM: 614556) and PPM1D (Jansen–de Vries syndrome, MIM: 617450) (16). We recommend thorough dysmorphic assessments with comprehensive HPO annotation, ideally performed by clinical geneticists before testing, to capture subtle yet diagnostically relevant traits and improve yield in CP and CP-like presentations. Beyond increasing diagnostic yield, reverse phenotyping also facilitated genotype–phenotype expansion, allowing the identification of novel or atypical clinical manifestations linked to known disease genes, thereby refining their phenotypic spectra.

. The third group comprised metabolic disorders, identified in four patients with variants in GATM (MIM: 612718), ACY1 (MIM: 609366), NDUFS3 (MIM: 606934), and NGLY1 (MIM: 615273). These genes are associated with treatable or manageable conditions that often present with CP-like features, including developmental delay, hypotonia, seizures, and progressive motor impairment. As highlighted in a systematic review by Leach et al., more than 50 inborn errors of metabolism can mimic CP, underscoring the value of early metabolic and genetic assessment to prevent irreversible neurological damage (41). Our findings support this, showing that integrating biochemical screening with genomic testing facilitates timely diagnosis, enables targeted interventions, and can significantly improve clinical outcomes in CP-like presentations.

The fourth group comprised patients with neuromuscular phenotypes, including congenital myopathies and hereditary motor neuropathies, with pathogenic variants in MFN2 (MIM: 608507), PLA2G6 (MIM: 603604), and COL6A1 (MIM: 120220) identified in four individuals. The patient (CP_P51.1) with an MFN2 variant presented with spastic paraparesis and brisk deep tendon reflexes, yet the overall presentation was more consistent with MFN2-related axonal neuropathy (MIM: 609260, 601,152). The PLA2G6 variant was detected in an individual (CP_P2.1) with motor delay, seizures, and structural brain anomalies, including corpus callosum hypoplasia and ventriculomegaly, features that overlap with, but are not typical of, CP. Other patients (CP_P8.1 and CP_P30.1) initially diagnosed with CP who exhibited neonatal hypotonia, delayed ambulation, and joint contractures were later found to harbor a COL6A1 variant, consistent with collagen VI-related muscular dystrophy (MIM: 158810, 254,090). These cases illustrate the phenotypic convergence between certain neuromuscular disorders and CP, particularly in the absence of perinatal risk factors or with atypical disease progression, and underscore the importance of genomic investigation for accurate diagnostic classification (42).

The fifth subgroup in our cohort encompassed hereditary spastic paraplegias (HSPs), a clinically and genetically heterogeneous group of neurodegenerative disorders defined by progressive lower-limb spasticity and gait impairment. Early-onset forms pose a particular diagnostic challenge, as hallmark motor features such as spasticity and delayed motor milestones substantially overlap with those of spastic Cerebral Palsy (CP), frequently resulting in initial misclassification (43). Among the HSP-associated genes, SPAST (MIM: 604277) and KIF1A (MIM: 601255) collectively account for over 40% of autosomal dominant HSP cases (44). SPAST encodes spastin, a microtubule-severing enzyme crucial for axonal growth, maintenance, and neuronal connectivity (45). Pathogenic SPAST variants demonstrate pronounced intrafamilial and interfamilial variability in age of onset, clinical severity, and disease trajectory, thereby complicating recognition (46). In our series, CP_P56.1 harbored pathogenic SPAST variants, each with a distinct phenotypic profile, despite sharing the same molecular diagnosis as the mother. KIF1A encodes an anterograde motor protein essential for the axonal transport of membranous organelles (47). One patient in our cohort, initially diagnosed with CP based solely on clinical presentation, was ultimately found to carry a pathogenic KIF1A variant. Notably, KIF1A-related disorders span a spectrum from HSP to complex neurodevelopmental phenotypes, including CP-like presentations with cognitive impairment and seizures (40). A defining distinction between HSP and CP lies in the progressive nature of spasticity in HSP, underscoring the necessity of genomic testing to accurately delineate these entities and guide prognosis, surveillance, and therapeutic strategies (48).

STRING network analysis showed a significant enrichment, indicating that the identified genes are not randomly distributed but functionally connected within shared biological modules. GO enrichment highlighted biological processes such as ion transport, membrane depolarization, and synaptic signaling, together with molecular functions related to voltage-gated ion channel activity and metal ion transport (49) (Supplementary Figure 2). Cellular component terms emphasized neuron projections, synapses, and ion channel complexes, supporting their localization to neuronal structures (Supplementary Figure 4). KEGG pathway analysis further revealed enrichment in cAMP, Ras, and thyroid hormone signaling, linking the candidate genes to key regulatory cascades (Supplementary Figure 5). Collectively, these convergent results suggest that disruption of ion homeostasis and neuronal excitability represents a central mechanism, consistent with disease enrichment findings that strongly associate these genes with epilepsy, autism spectrum disorder, and intellectual disability.

When interpreting our results, several limitations should be considered. The use of heterogeneous sequencing platforms (CES, WES, and WGS) affected the resolution for detecting certain variant types, particularly copy number variants and mitochondrial DNA variants. The relatively small WGS subset further limited genome-wide variant detection, while CNV analysis was constrained by the resolution of WES data. In addition, clinical and phenotypic data were collected retrospectively, and Human Phenotype Ontology (HPO) terms were extracted from hospital-based records, which may have resulted in incomplete phenotypic annotation and limited genotype–phenotype correlations. The enrichment of genetically suspected cases may also have introduced selection bias. Finally, the modest cohort size limits the generalizability of our findings, underscoring the need for larger, multicenter studies integrating trio-based WES/WGS, high-resolution CNV detection, deep phenotyping, and functional validation of clinically relevant variants to improve diagnostic yield and precision.

In conclusion, our findings underscore the pivotal role of genomic testing in refining the diagnosis of CP and CP-like presentations, revealing a broad spectrum of underlying genetic etiologies that often extend beyond traditional clinical boundaries. In line with the International Cerebral Palsy Genomics Consortium (ICPGC) consensus, when a pathogenic or likely pathogenic variant is identified in a patient whose phenotype fulfills the consensus definition of CP, the diagnosis should be retained rather than reclassified as a mimic, highlighting the need to integrate genetic findings into, rather than replace, clinical classification. By integrating advanced sequencing technologies, reverse phenotyping, and deep clinical characterization, we not only increased diagnostic yield but also expanded the known genotype–phenotype spectrum of several disease genes, including many not previously associated with CP. These results highlight the importance of considering genomic evaluation early in the diagnostic pathway, particularly for patients with atypical features or without clear perinatal risk factors, to enable timely and targeted interventions. Moving forward, larger multicenter studies employing standardized trio-based WES/WGS, high-resolution CNV analysis, and systematic phenotyping will be essential for fully delineating the genetic architecture of CP, improving diagnostic precision, and ultimately guiding personalized management strategies for affected individuals.

 Data availability statement

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding authors.

 Ethics statement

The studies involving humans were approved by the Acıbadem Healthcare Institutions Medical Research Ethics Committee (ATADEK) under decision number 2022–04/18. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin.

 Author contributions

AY: Methodology, Visualization, Conceptualization, Validation, Writing – original draft, Formal analysis, Investigation, Writing – review & editing, Data curation. OA-D: Writing – original draft, Investigation, Conceptualization, Validation, Supervision, Project administration, Methodology, Writing – review & editing. ZO: Writing – review & editing, Data curation, Visualization, Formal analysis, Writing – original draft. GB: Conceptualization, Writing – review & editing, Investigation, Supervision, Writing – original draft, Funding acquisition, Data curation, Project administration. SA: Validation, Writing – review & editing, Investigation, Resources, Writing – original draft. PG: Writing – original draft, Resources, Validation, Writing – review & editing, Investigation. HM: Validation, Writing – review & editing, Writing – original draft, Investigation, Resources. BKu: Investigation, Writing – original draft, Resources, Validation, Writing – review & editing. BKa: Writing – review & editing, Writing – original draft, Resources, Validation, Investigation. HS: Investigation, Writing – original draft, Writing – review & editing, Resources, Validation. NA: Writing – review & editing, Writing – original draft, Investigation, Resources, Validation. OO: Software, Investigation, Writing – original draft, Methodology, Writing – review & editing, Supervision. KB: Project administration, Validation, Methodology, Writing – original draft, Supervision, Formal analysis, Resources, Writing – review & editing, Conceptualization, Investigation, Funding acquisition. UO: Supervision, Writing – review & editing, Project administration, Funding acquisition, Investigation, Writing – original draft, Resources, Conceptualization.

 Acknowledgments

The authors also acknowledge the support of the RareBoost Project, funded by the European Union’s Horizon 2020 Research and Innovation Programme (Grant No. 952346, https://rareboost.ibg.edu.tr/), and the İSTisNA Project (Istanbul Undiagnosed and Rare Diseases Solution Platform, Project No. TR10/19/FZD/0003, https://istisna.org/en/about-us/), both of which contributed to research capacity and collaborative infrastructure in the field of rare diseases. We are grateful to the participants and their families for their support of this research.

 Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

 Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

 Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fneur.2026.1742186/full#supplementary-material

 References JinSC LewisSA BakhtiariS ZengX SierantMC ShettyS . Mutations disrupting neuritogenesis genes confer risk for cerebral palsy. Nat Genet. (2020) 52:1046–56. doi: 10.1038/s41588-020-0695-1, 32989326 Smithers-SheedyH BadawiN BlairE CansC HimmelmannK Krageloh-MannI . What constitutes cerebral palsy in the twenty-first century? Dev Med Child Neurol. (2014) 56:323–8. doi: 10.1111/dmcn.12262, 24111874 Surveillance of Cerebral Palsy in Europe. Surveillance of cerebral palsy in Europe: a collaboration of cerebral palsy surveys and registers. Surveillance of cerebral palsy in Europe (SCPE). Dev Med Child Neurol. (2000) 42:816–24. doi: 10.1111/j.1469-8749.2000.tb00695.x KorzeniewskiSJ SlaughterJ LenskiM HaakP PanethN. The complex aetiology of cerebral palsy. Nat Rev Neurol. (2018) 14:528–43. doi: 10.1038/s41582-018-0043-6, 30104744 DelacyMJ ReidSM Australian Cerebral Palsy RegisterG. Profile of associated impairments at age 5 years in Australia by cerebral palsy subtype and Gross motor function classification system level for birth years 1996 to 2005. Dev Med Child Neurol. (2016) 58:50–6. doi: 10.1111/dmcn.13012 McIntyreS GoldsmithS WebbA EhlingerV HollungSJ McConnellK . Global prevalence of cerebral palsy: a systematic analysis. Dev Med Child Neurol. (2022) 64:1494–506. doi: 10.1111/dmcn.15346, 35952356 XuY LiY RichardSA SunY ZhuC. Genetic pathways in cerebral palsy: a review of the implications for precision diagnosis and understanding disease mechanisms. Neural Regen Res. (2024) 19:1499–508. doi: 10.4103/1673-5374.385855, 38051892 PaulS NaharA BhagawatiM KunwarAJ. A review on recent advances of cerebral palsy. Oxidative Med Cell Longev. (2022) 2022:2622310. doi: 10.1155/2022/2622310, 35941906 Abd ElmagidDS MagdyH. Evaluation of risk factors for cerebral palsy. Egypt J Neurol Psychiatry Neurosurg. (2021) 57:13. doi: 10.1186/s41983-020-00265-1 FehlingsDL ZarreiM EngchuanW SondheimerN ThiruvahindrapuramB MacDonaldJR . Comprehensive whole-genome sequence analyses provide insights into the genomic architecture of cerebral palsy. Nat Genet. (2024) 56:585–94. doi: 10.1038/s41588-024-01686-x SrivastavaS LewisSA CohenJS ZhangB AravamuthanBR ChopraM . Molecular diagnostic yield of exome sequencing and chromosomal microarray in Cerebral Palsy: a systematic review and Meta-analysis. JAMA Neurol. (2022) 79:1287–95. doi: 10.1001/jamaneurol.2022.3549, 36279113 ThysL BeysenD CeulemansB KenisS DielmanC RoelensF . The genetic puzzle of Cerebral Palsy: results of a monocentric study. Pediatr Neurol. (2024) 161:1–8. doi: 10.1016/j.pediatrneurol.2024.07.019, 39213953 ChopraM GableDL Love-NicholsJ TsaoA RockowitzS SlizP . Mendelian etiologies identified with whole exome sequencing in cerebral palsy. Ann Clin Transl Neurol. (2022) 9:193–205. doi: 10.1002/acn3.51506, 35076175 ReuterCM KohlerJN BonnerD ZastrowD FernandezL DriesA . Yield of whole exome sequencing in undiagnosed patients facing insurance coverage barriers to genetic testing. J Genet Couns. (2019) 28:1107–18. doi: 10.1002/jgc4.1161, 31478310 van EykCL WebberDL MinocheAE Perez-JuradoLA CorbettMA GardnerAE . Yield of clinically reportable genetic variants in unselected cerebral palsy by whole genome sequencing. NPJ Genom Med. (2021) 6:74. doi: 10.1038/s41525-021-00238-0, 34531397 YechieliM GulsunerS Ben-PaziH FattalA AranA KuzminskyA . Diagnostic yield of chromosomal microarray and trio whole exome sequencing in cryptogenic cerebral palsy. J Med Genet. (2022) 59:759–67. doi: 10.1136/jmedgenet-2021-107884, 34321325 EwansLJ SchofieldD ShresthaR ZhuY GayevskiyV YingK . Whole-exome sequencing reanalysis at 12 months boosts diagnosis and is cost-effective when applied early in Mendelian disorders. Genet Med. (2018) 20:1564–74. doi: 10.1038/gim.2018.39, 29595814 BullichG MatalongaL PujadasM PapakonstantinouA PisciaD TondaR . Systematic collaborative reanalysis of genomic data improves diagnostic yield in neurologic rare diseases. J Mol Diagn. (2022) 24:529–42. doi: 10.1016/j.jmoldx.2022.02.003, 35569879 LiuP MengL NormandEA XiaF SongX GhaziA . Reanalysis of clinical exome sequencing data. N Engl J Med. (2019) 380:2478–80. doi: 10.1056/NEJMc1812033, 31216405 BakerSW MurrellJR NesbittAI PechterKB BalciunieneJ ZhaoX . Automated clinical exome reanalysis reveals novel diagnoses. J Mol Diagn. (2019) 21:38–48. doi: 10.1016/j.jmoldx.2018.07.008, 30577886 World Medical Association. World medical association declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. (2013) 310:2191–4. doi: 10.1001/jama.2013.281053 Geniva Informatics. Gennext workflows [computer software]. Zenodo; (2025). doi: 10.5281/zenodo.17426020 WjbMZ CurtisK FoxA PattersonD ShenkerS StoicaI . Faster and more accurate sequence alignment with SNAP. arXiv. (2011) 1111.5572v1 HerzeelC CostanzaP DecapD FostierJ ReumersJ. elPrep: high-performance preparation of sequence alignment/map files for variant calling. PLoS One. (2015) 10:e0132868. doi: 10.1371/journal.pone.0132868, 26182406 PoplinR Ruano-RubioV DePristoMA FennellTJ CarneiroMO AuweraGAV . Scaling accurate genetic variant discovery to tens of thousands of samples. bioRxiv. (2018) 201178 PoplinR ChangPC AlexanderD SchwartzS ColthurstT KuA . A universal SNP and small-indel variant caller using deep neural networks. Nat Biotechnol. (2018) 36:983–7. doi: 10.1038/nbt.4235, 30247488 KarczewskiKJ FrancioliLC TiaoG CummingsBB AlfoldiJ WangQ . The mutational constraint spectrum quantified from variation in 141,456 humans. Nature. (2020) 581:434–43. doi: 10.1038/s41586-020-2308-7, 32461654 ChenS FrancioliLC GoodrichJK CollinsRL KanaiM WangQ . A genome-wide mutational constraint map quantified from variation in 76,156 human genomes. bioRxiv. (2022) 2022.03.20.485034 KarsME BasakAN OnatOE BilguvarK ChoiJ ItanY . The genetic structure of the Turkish population reveals high levels of variation and admixture. Proc Natl Acad Sci USA. (2021) 118:e2120031118. doi: 10.1073/pnas.2026076118 RentzschP WittenD CooperGM ShendureJ KircherM. CADD: predicting the deleteriousness of variants throughout the human genome. Nucleic Acids Res. (2019) 47:D886–94. doi: 10.1093/nar/gky1016, 30371827 AdzhubeiI JordanDM SunyaevSR. Predicting functional effect of human missense mutations using PolyPhen-2. Curr Protoc Hum Genet. (2013) 76:Unit7 20. doi: 10.1002/0471142905.hg0720s76 KimS JhongJH LeeJ KooJY. Meta-analytic support vector machine for integrating multiple omics data. BioData Min. (2017) 10:2. doi: 10.1186/s13040-017-0126-8, 28149325 NgPC HenikoffS. SIFT: predicting amino acid changes that affect protein function. Nucleic Acids Res. (2003) 31:3812–4. doi: 10.1093/nar/gkg509, 12824425 JaganathanK Kyriazopoulou PanagiotopoulouS McRaeJF DarbandiSF KnowlesD LiYI . Predicting splicing from primary sequence with deep learning. Cell. (2019) 176:535–548.e24. doi: 10.1016/j.cell.2018.12.015, 30661751 RichardsS AzizN BaleS BickD DasS Gastier-FosterJ . Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. (2015) 17:405–24. doi: 10.1038/gim.2015.30, 25741868 Moreno-De-LucaA MillanF PesacretaDR ElloumiHZ OetjensMT TeigenC . Molecular diagnostic yield of exome sequencing in patients with cerebral palsy. JAMA. (2021) 325:467–75. doi: 10.1001/jama.2020.26148, 33528536 van EykCL FaheyMC GeczJ. Redefining cerebral palsies as a diverse group of neurodevelopmental disorders with genetic aetiology. Nat Rev Neurol. (2023) 19:542–55. doi: 10.1038/s41582-023-00847-6, 37537278 Akgun-DoganO Tuc BengurE AyB OzkoseGS KarE BengurFB . Impact of deep phenotyping: high diagnostic yield in a diverse pediatric population of 172 patients through clinical whole-genome sequencing at a single center. Front Genet. (2024) 15:1347474. doi: 10.3389/fgene.2024.1347474 NollR SchaafJ Von WagnerM StorfH. Diagnosis support for rare diseases using phenotypic profiles. Stud Health Technol Inform. (2023) 302:607–8. doi: 10.3233/SHTI230216, 37203759 HakamiWS HundallahKJ TabarkiBM. Metabolic and genetic disorders mimicking cerebral palsy. Neurosciences. (2019) 24:155–63. doi: 10.17712/nsj.2019.3.20190045, 31380813 LeachEL ShevellM BowdenK Stockler-IpsirogluS van KarnebeekCD. Treatable inborn errors of metabolism presenting as cerebral palsy mimics: systematic literature review. Orphanet J Rare Dis. (2014) 9:197. doi: 10.1186/s13023-014-0197-2, 25433678 ClewesK HammondC DongY MeyerM LoweE RoseJ. Neuromuscular impairments of cerebral palsy: contributions to gait abnormalities and implications for treatment. Front Hum Neurosci. (2024) 18:1445793. doi: 10.3389/fnhum.2024.1445793, 39359619 FaheyMC MaclennanAH KretzschmarD GeczJ KruerMC. The genetic basis of cerebral palsy. Dev Med Child Neurol. (2017) 59:462–9. doi: 10.1111/dmcn.13363, 28042670 WilsonYA GarrityN Smithers-SheedyH GoldsmithS KarimT HenryG . Clinically relevant genes identified in cerebral palsy cohorts following evaluation of the clinical description and phenotype: a systematic review. J Child Neurol. (2024) 39:500–9. doi: 10.1177/08830738241277231, 39246294 JeongB KimTH KimDS ShinWH LeeJR KimNS . Spastin contributes to neural development through the regulation of microtubule dynamics in the primary cilia of neural stem cells. Neuroscience. (2019) 411:76–85. doi: 10.1016/j.neuroscience.2019.05.024, 31150727 YaoL CaoY ZhangC HuangX TianW CaoL. Clinical and genetic characteristics in a Chinese cohort of complex spastic paraplegia type 4. Clin Genet. (2024) 106:56–65. doi: 10.1111/cge.14510, 38403837 ChibaK KitaT AnazawaY NiwaS. Insight into the regulation of axonal transport from the study of KIF1A-associated neurological disorder. J Cell Sci. (2023) 136:jcs260742. doi: 10.1242/jcs.260742 BasuAP LowK RatnaikeT RowitchD. Genetic investigations in cerebral palsy. Dev Med Child Neurol. (2025) 67:177–85. doi: 10.1111/dmcn.16080, 39208295 AshburnerM BallCA BlakeJA BotsteinD ButlerH CherryJM . Gene ontology: tool for the unification of biology. Nat Genet. (2000) 25:25–9. doi: 10.1038/7555

Edited by: Sheng Luo, The Second Affiliated Hospital of Guangzhou Medical University, China

Reviewed by: Maria Cristina Aspromonte, University of Padua, Italy

Peng-Yu Wang, Second Affiliated Hospital of Guangzhou Medical University, China

https://github.com/GenivaInformatics/gennext-workflows

https://panelapp-aus.org/panels/73/

https://string-db.org/

http://www.genome.jp/kegg