Intellectual disability syndromic and non-syndromic
Gene: CSMD2 Red List (low evidence)Red List (low evidence)
PMID: 40632521 Li et al 2025 (Epilepsia) reported 6 unrelated individuals of Han Chinese descent with biallelic CSMD2 missense variants (NM_052896) and focal epilepsy. 5 individuals were compound heterozygous and one was homozygous. These individuals were ascertained through trio WES analysis of 420 unrelated individuals with focal epilepsy enrolled in the China Epilepsy Gene 1.0 project.
Phenotypic features
- age of onset 1.5-10 years old
- complex partial seizures (4), secondary GTCS (2)
- Normal MRI-B (3), focal cortical dysplasia (1)
- mild ID (1).
The variants were noted to be rare in EXAC-East Asian cohort, most located in CUB/Sushi domains. The gene has some evidence of missense and LoF constraint in gnomAD v4. There was also enrichment of biallelic CSMD2 variants in affected individuals versus a control cohort of unaffected parents (5/420 compound hets affected individuals, 3/1942 compound hets in unaffected parents). Previous mouse Csmd2 knockdown models demonstrated reduction in dendritic spine density and complexity. LoF is the postulated disease mechanism.
Closely related gene paralog CSMD1 has a definitive association with autosomal recessive complex neurodevelopmental disorder with a more severe phenotype. Different expression profiles during developmental stages between CSMD1 and CSMD2 postulated for the comparatively milder phenotype associated with the latter.
CSMD2 has 71 exons and 3631 amino acids. The true prevalence of biallelic missense variants in healthy individuals across diverse ancestries has not been ascertained. Review of the missense variants in this study highlighted issues in a number of them including poor-moderate conservation, conflicting or benign in silicos including REVEL, non-coding in an alternative transcript, Case 4 p.Val1547Ile homozygote – this variant has been noted in an East Asian male homozygote aged between 45-50 in gnomAD v4. In addition, no information about unaffected/affected siblings and segregation testing has been provided
Given prevalence of focal epilepsy, stronger case-control evidence from diverse ancestries and variant-specific functional evidence is required to support this proposed gene-disease association.
Sources: LiteratureCreated: 4 Sep 2025, 12:03 a.m. | Last Modified: 4 Sep 2025, 4:56 a.m.
Panel Version: 1.273
Mode of inheritance
BIALLELIC, autosomal or pseudoautosomal
Phenotypes
Focal epilepsy - MONDO:0005384, CSMD2-related
Publications
Gene: csmd2 has been classified as Red List (Low Evidence).
gene: CSMD2 was added gene: CSMD2 was added to Intellectual disability syndromic and non-syndromic. Sources: Literature Mode of inheritance for gene: CSMD2 was set to BIALLELIC, autosomal or pseudoautosomal Publications for gene: CSMD2 were set to PMID: 40632521; 31068362; 38649688 Phenotypes for gene: CSMD2 were set to Focal epilepsy - MONDO:0005384, CSMD2-related Review for gene: CSMD2 was set to RED
If promoting or demoting a gene, please provide comments to justify a decision to move it.
Genes included in a Genomics England gene panel for a rare disease category (green list) should fit the criteria A-E outlined below.
These guidelines were developed as a combination of the ClinGen DEFINITIVE evidence for a causal role of the gene in the disease(a), and the Developmental Disorder Genotype-Phenotype (DDG2P) CONFIRMED DD Gene evidence level(b) (please see the original references provided below for full details). These help provide a guideline for expert reviewers when assessing whether a gene should be on the green or the red list of a panel.
A. There are plausible disease-causing mutations(i) within, affecting or encompassing an interpretable functional region(ii) of this gene identified in multiple (>3) unrelated cases/families with the phenotype(iii).
OR
B. There are plausible disease-causing mutations(i) within, affecting or encompassing cis-regulatory elements convincingly affecting the expression of a single gene identified in multiple (>3) unrelated cases/families with the phenotype(iii).
OR
C. As definitions A or B but in 2 or 3 unrelated cases/families with the phenotype, with the addition of convincing bioinformatic or functional evidence of causation e.g. known inborn error of metabolism with mutation in orthologous gene which is known to have the relevant deficient enzymatic activity in other species; existence of an animal model which recapitulates the human phenotype.
AND
D. Evidence indicates that disease-causing mutations follow a Mendelian pattern of causation appropriate for reporting in a diagnostic setting(iv).
AND
E. No convincing evidence exists or has emerged that contradicts the role of the gene in the specified phenotype.
(i)Plausible disease-causing mutations: Recurrent de novo mutations convincingly affecting gene function. Rare, fully-penetrant mutations - relevant genotype never, or very rarely, seen in controls. (ii) Interpretable functional region: ORF in protein coding genes miRNA stem or loop. (iii) Phenotype: the rare disease category, as described in the eligibility statement. (iv) Intermediate penetrance genes should not be included.
It’s assumed that loss-of-function variants in this gene can cause the disease/phenotype unless an exception to this rule is known. We would like to collect information regarding exceptions. An example exception is the PCSK9 gene, where loss-of-function variants are not relevant for a hypercholesterolemia phenotype as they are associated with increased LDL-cholesterol uptake via LDLR (PMID: 25911073).
If a curated set of known-pathogenic variants is available for this gene-phenotype, please contact us at panelapp@genomicsengland.co.uk
We classify loss-of-function variants as those with the following Sequence Ontology (SO) terms:
Term descriptions can be found on the PanelApp homepage and Ensembl.
If you are submitting this evaluation on behalf of a clinical laboratory please indicate whether you report variants in this gene as part of your current diagnostic practice by checking the box
Standardised terms were used to represent the gene-disease mode of inheritance, and were mapped to commonly used terms from the different sources. Below each of the terms is described, along with the equivalent commonly-used terms.
A variant on one allele of this gene can cause the disease, and imprinting has not been implicated.
A variant on the paternally-inherited allele of this gene can cause the disease, if the alternate allele is imprinted (function muted).
A variant on the maternally-inherited allele of this gene can cause the disease, if the alternate allele is imprinted (function muted).
A variant on one allele of this gene can cause the disease. This is the default used for autosomal dominant mode of inheritance where no knowledge of the imprinting status of the gene required to cause the disease is known. Mapped to the following commonly used terms from different sources: autosomal dominant, dominant, AD, DOMINANT.
A variant on both alleles of this gene is required to cause the disease. Mapped to the following commonly used terms from different sources: autosomal recessive, recessive, AR, RECESSIVE.
The disease can be caused by a variant on one or both alleles of this gene. Mapped to the following commonly used terms from different sources: autosomal recessive or autosomal dominant, recessive or dominant, AR/AD, AD/AR, DOMINANT/RECESSIVE, RECESSIVE/DOMINANT.
A variant on one allele of this gene can cause the disease, however a variant on both alleles of this gene can result in a more severe form of the disease/phenotype.
A variant in this gene can cause the disease in males as they have one X-chromosome allele, whereas a variant on both X-chromosome alleles is required to cause the disease in females. Mapped to the following commonly used term from different sources: X-linked recessive.
A variant in this gene can cause the disease in males as they have one X-chromosome allele. A variant on one allele of this gene may also cause the disease in females, though the disease/phenotype may be less severe and may have a later-onset than is seen in males. X-linked inactivation and mosaicism in different tissues complicate whether a female presents with the disease, and can change over their lifetime. This term is the default setting used for X-linked genes, where it is not known definitately whether females require a variant on each allele of this gene in order to be affected. Mapped to the following commonly used terms from different sources: X-linked dominant, x-linked, X-LINKED, X-linked.
The gene is in the mitochondrial genome and variants within this can cause this disease, maternally inherited. Mapped to the following commonly used term from different sources: Mitochondrial.
Mapped to the following commonly used terms from different sources: Unknown, NA, information not provided.
For example, if the mode of inheritance is digenic, please indicate this in the comments and which other gene is involved.