The CACNA1A Cure Pathway

The six projects within this series address therapeutic design at different points in the cure pipeline. They are organized in rough order of clinical maturity: from the most mature therapeutic strategy (ASO silencing, Layer 1) toward the most foundational infrastructure (modifier gene prediction, Layer 6). 

AlleleSelect (Layer 1)

• AlleleSelect designs allele-selective antisense oligonucleotides (ASOs) targeting CACNA1A gain-of-function mutations. ASOs are short molecular sequences that bind to the mutant mRNA produced by the broken gene copy and mark it for degradation, while the healthy copy continues producing functional CaV2.1 channels. The central challenge is selectivity: the mutant and healthy mRNAs often differ by a single nucleotide, so the design has to be precise enough to distinguish them. AlleleSelect computes this selectivity using thermodynamic binding energy differences and coevolutionary position scoring, ranks candidates by predicted allele-selective activity, and includes an integrated delivery assessment module that evaluates CNS feasibility for each candidate.

VectoSelect (Layer 2)

• Gene therapy (delivering a working copy of CACNA1A directly into neurons via a viral vector) is one of the most powerful therapeutic approaches for dominant channelopathies, but clinical programs have to choose which variants to target in early preclinical studies. Viral vectors have cargo size limits, and research resources are finite. VectoSelect integrates clinical severity data, biophysical variant characterization, existing mouse model availability, and gene therapy tractability scores to rank CACNA1A variants by their priority as therapeutic targets. It was built specifically to support helper-dependent adenoviral (HdAd) vector work, which has demonstrated Purkinje cell transduction relevant to the cerebellar phenotypes of CACNA1A disease.

PrimeSelect (Layer 3)

• Prime editing uses a modified CRISPR system to find a specific genomic location and rewrite a small stretch of sequence — correcting a single wrong nucleotide without cutting both DNA strands. For CACNA1A missense variants, this requires designing a pegRNA: a guide that directs the editor to the right location and a template that carries the corrected sequence. PrimeSelect automates this design for known CACNA1A missense variants, generating epegRNA candidates ranked by predicted editing efficiency, off-target risk, and CNS delivery feasibility.

CompoundSelect (Layer 4)

• Before a new drug is developed, existing approved drugs should be systematically evaluated for repurposing. CompoundSelect takes a CACNA1A variant and its functional class (GOF or LOF) and queries FDA-approved compound databases filtered for CNS penetrance, scoring each compound by its predicted mechanism match, blood-brain barrier penetration, safety profile, and CaV2.1 specificity. A September 2025 paper demonstrated this approach for cardiac channelopathy; CompoundSelect extends the paradigm computationally for CACNA1A, where the BBB constraint dramatically narrows the viable candidate pool.

BiomarkSelect (Layer 5)

• Clinical trials need measurable outcomes. The CACNA1A Foundation's 2023 Research Roundtable identified explicitly that no validated biomarkers exist for CACNA1A disease- a gap that blocks trial design regardless of how promising a drug candidate is. BiomarkSelect takes a CACNA1A variant and its phenotype tier and predicts which measurable clinical endpoints  (EEG features, cerebellar atrophy on MRI, clinical ataxia scales, migraine-specific measures, cognitive assessments) are most likely to be informative for tracking treatment response in that patient. The N-acetyl-L-leucine Phase III trial currently enrolling needs exactly this kind of variant-stratified biomarker guidance.

BiomarkSelect (Layer 5)

• Two siblings with the same CACNA1A variant can have dramatically different clinical outcomes. The variant alone does not fully explain this. ModifierSelect predicts which secondary variants in genes encoding proteins that interact with or regulate CaV2.1 (the β subunit genes (CACNB1-4), α2δ subunit genes, related calcium channels, seizure threshold genes) would worsen or ameliorate the CACNA1A phenotype, using gnomAD population data and published functional interaction evidence. This addresses a problem Dr. Noebels identified in his 2021 roadmap talk and that a 2024 Frontiers in Neurology paper on intrafamilial CACNA1A variability called for explicitly.

The CACNA1A Cure Pathway is a set of six computational tools designed to speed up research toward treatments for disorders caused by mutations in the CACNA1A gene. These disorders are mostly dominant, meaning a patient inherits one broken copy of the gene and one healthy copy. The goal of any treatment, then, is to silence the broken copy, replace it with a working one, correct the mutation directly, or find existing drugs that compensate for what has gone wrong. Currently, no cure exists, and researchers working on these approaches often have to start from scratch. They manually design potential drugs, guess which variants to study first, or reinvent basic steps. This series attempts to build a computational layer above each strategy so that scientists can move faster and with more precision.

The six tools are organized roughly by how close each therapeutic approach is to clinical use. AlleleSelect designs allele-selective antisense oligonucleotides (ASOs), which are tiny molecules that can find and mark the broken copy of the gene for destruction while leaving the healthy copy alone. The main challenge is selectivity. Often the only difference between the mutant and healthy message is a single chemical letter, so AlleleSelect uses binding energy calculations to rank the most promising candidates. VectoSelect helps prioritize which variants to target first when developing gene therapy, since viral vectors have size limits and research budgets are finite. It ranks variants by clinical severity, existing mouse models, and how easy they are to target. PrimeSelect automates the design of prime editing guide RNAs (pegRNAs), a precise CRISPR based tool that rewrites a single wrong DNA letter without cutting both strands. CompoundSelect screens existing FDA approved drugs for possible repurposing, with special attention to whether a drug can cross the blood brain barrier. This is a major constraint for CACNA1A. BiomarkSelect predicts which clinical measurements (EEG features, MRI atrophy scores, ataxia scales) would best track treatment response in a given patient, because clinical trials cannot even start without knowing what to measure. Finally, ModifierSelect addresses a puzzle: two siblings with the same CACNA1A mutation can have very different outcomes. This tool looks for secondary variants in interacting genes (like calcium channel subunits or seizure threshold genes) that might worsen or ameliorate the disease.