From Nucleotide Change to Neuronal Dysfunction: Computational Interpretation of Coding Variants in Neurogenetics

Variants can be assigned to classes of variants based on the molecular consequence of a nucleotide change, e.g. substitution in a protein-coding sequence (PCG). Substitutions in a PCG can result in several outcomes: the encoded amino acid can be preserved, it can be replaced by another amino acid, or a premature termination codon can be introduced. These variants are commonly classified as synonymous, missense, and nonsense variants, respectively. However, classifying a variant into one of these categories does not necessarily reveal whether the variant has an effect on the protein or even causes a disease.
Here we review coding variation and discuss how analysis of DNA sequences can help to predict the molecular consequences of variants. Synonymous variation can affect RNA splicing, mRNA stability, translation and protein folding (Sauna and Kimchi-Sarfaty 2011; Hunt et al. 2014). Such variation can have a significant impact on protein function despite the coding DNA sequence encoding the same amino acid(s) as in wild-type sequence. A large number of missense variants have been studied to predict their potential impact on protein function and these variants can affect protein function in a number of ways including protein stability, channel function (e.g. in ion channel proteins), protein–protein interactions and protein function within the cell (Ng and Henikoff 2001; Adzhubei et al. 2010). The impact of any given missense variant will depend on a number of factors including the biochemical and structural properties of the affected amino acid. Nonsense variants result in the translation of a truncated protein and can result in the degradation of the nonsense-containing mRNA by nonsense-mediated mRNA decay (Lindeboom, Supek, and Lehner 2016). Here, we review the evidence for the functional consequences of variants in SCN1A, a gene whose products are responsible for neuronal sodium channel function. While variants in coding regions can affect protein function in a number of ways, they can be broadly classified into three types: synonymous variation; missense variants; and nonsense variants. Such variants have been found to affect protein function in a number of ways and their functional consequences can be highly variable. Functional studies of variants in SCN1A have found that variants that have the same broad classification can have very different consequences for channel function (Brunklaus et al. 2020; Knox et al. 2025). More recently, four variants that affect the same codon in SCN1A have been found to have very different functional consequences for sodium channel function (Liebovitz et al. 2026). The functional studies were carried out using cDNA that encoded the variants of interest and the variants were expressed in a cell line that endogenously expressed sodium channels. In each case, the variants resulted in a change in the protein sequence that was determined by the cDNA sequence. The four variants resulted in different changes to channel function, each of which highlighted an important feature of sodium channel function. In each case, a change to sodium channel function could have a profound impact on neuronal function, particularly in neurons that required high frequency firing. As functional studies of all coding variants of interest would not be practical, it would be highly desirable to identify variants of interest that had a high likelihood of having a functional consequence. Here, we review how analysis of cDNA and coding DNA sequence can help to predict the consequences of any coding variant of interest. In the case of the four variants of interest that affected the same codon in SCN1A, prediction would have indicated that each of the variants was of interest. However, no variant would have predicted the extent of the change to sodium channel function that was found in functional studies. Thus, while prediction of coding variants of interest is highly desirable, such studies would need to be underpinned by functional studies to establish the full extent of the consequences of any coding variant of interest.
. These findings indicate that computational sequence classification is most useful as an initial annotation step. More complete interpretation requires population data, evolutionary conservation, structural information, functional experiments and clinical evidence.