01/3D Structure
? About the 3D Viewer
Mol* (pronounced "molstar") is an open-source molecular visualization tool used by the Protein Data Bank and AlphaFold Database. Learn more at molstar.org.
Controls:
- Rotate: Click and drag
- Zoom: Scroll wheel or pinch
- Pan: Right-click and drag (or two-finger drag)
- Reset: Double-click to reset view
What am I looking at?
This is a predicted 3D structure of the protein. The ribbon diagram shows the protein backbone—helices appear as coils, sheets as arrows, and loops as simple lines. The shape determines how the protein functions: where it binds to other molecules, how it catalyzes reactions, and how mutations might disrupt its activity.
Color legend:
The structure is colored by pLDDT confidence score, which indicates how confident AlphaFold is in each region's predicted position:
- Blue (>90): Very high confidence
- Cyan (70-90): Confident
- Yellow (50-70): Low confidence
- Orange (<50): Very low confidence, likely disordered
02/AI Analysis
TLDR
Spinocerebellar ataxia type 3 (also called Machado-Joseph disease) is caused by an abnormal expansion of CAG repeats in the ATXN3 gene, which produces a protein with too many glutamine amino acids that forms toxic clumps in brain cells. This structural prediction of the normal ATXN3 protein with 27 glutamines (Q27) achieved moderate confidence (average score 71.9 out of 100), revealing both well-defined functional regions and flexible domains that may be critical for understanding how the expanded versions cause disease. While the moderate confidence limits definitive structural conclusions, this model provides a foundation for comparing how disease-causing expansions might alter protein behavior and aggregation patterns.
Detailed Analysis
Works Cited
Similar Research
03/Research Data
ClinVar Classification
Not found in ClinVar
Population Frequency
No population data available
Disease Associations
875 totalShowing 5 of 875 associations
AI Research Brief
04/AlphaFold Metrics
No visualization images available.
05/Domain Annotations
Structural Domains & Regions
Functional Sites
Binding Partners
Gene Ontology
06/Structural Caption
ATXN3 wild-type (Q27) shows well-folded Josephin domain and three UIM motifs, with expected disorder in the central linker region (residues 258-338).
Average pLDDT of 71.9 with 63% high-confidence residues (229/361). The disordered region (residues 258-338) and C-terminal segments show reduced confidence, while the Josephin domain exhibits higher structural confidence.
The N-terminal Josephin domain (residues 1-180) corresponds to high-confidence structured regions. UIM motifs show moderate confidence, with UIM 1 and 2 (residues 224-263) better predicted than UIM 3 (residues 331-349). The annotated disordered region (residues 258-338) aligns with low-confidence predictions, consistent with intrinsic disorder.
Wild-type fold with Q27 polyglutamine tract — no variant mutation. This represents the baseline ATXN3 structure with normal-length polyQ expansion in the Josephin domain.
07/Peptide Therapeutics
Aggregation Analysis
Aggregation propensity analysis identifies 1 hotspots (average score: -0.00) using Pawar+KyteDoolittle+charge algorithm.
08/Known Inhibitors
No known inhibitors found. Run peptide agent to search literature.
09/Candidate Peptides
De Novo Peptide Design Pipeline
Pipeline: BoltzGen (de novo binder design) → Boltz-2 rescore → 8-gate wetlab filter → PK + BBB advisory gates. Target site selected from UniProt curated annotations, P2Rank pocket prediction, and aggregation propensity (in that priority order). Advisory gates annotate each candidate with estimated serum half-life, renal/immunogenicity risk, and (for CNS targets) a recommended blood-brain-barrier shuttle conjugation — without silently dropping designs.
Loading candidate statistics...
Sequences are withheld pending IP review. Full candidate data (sequences,
scores, CIF files) is available to authorized reviewers via the
/api/private/candidates/{fold_id} endpoint with
X-Private-Key.
Legacy candidates (charge-complementary)
Target Region
Residues 148–152 (0.56 aggregation score)Candidate ID
CP-ATXN3-001
(7 residues · computational design)
10/Agent Findings
Literature Agent (1)
These papers are highly relevant for understanding ATXN3 wildtype Q27 function in the context of SCA3 pathogenesis. They provide critical insights into the biophysical properties of normal versus expanded polyQ ATXN3 (phase separation, aggregation), cellular stress response mechanisms disrupted by mutation, genetic modifiers that influence disease phenotype, and longitudinal biomarkers for tracking disease progression. Understanding wildtype ATXN3 behavior is essential for developing targeted therapies and interpreting how normal-range CAG repeats (like Q27) differ mechanistically from pathogenic expansions.
Clinical Agent (1)
The collection of first baseline data for the ATXN3 wildtype Q27 variant establishes a normal reference point, as Q27 (27 CAG repeats) falls within the non-pathogenic range for Machado-Joseph disease (pathogenic expansions typically begin at ≥60 repeats). This baseline is clinically significant for comparing against pathogenic expanded alleles to quantify the relationship between polyglutamine tract length and disease penetrance, age of onset, and symptom severity. These data enable researchers to distinguish normal ATXN3 function from the toxic gain-of-function mechanism associated with expanded repeats, which is essential for developing length-dependent therapeutic strategies.
Structural Agent (1)
AlphaFold structure update: Baseline check: 5 structure(s) found
Supplements Agent (1)
The current research landscape shows no active investigation of supplement or peptide-based therapeutics specifically targeting wild-type ATXN3 (Q27) in Spinocerebellar ataxia type 3. The available literature focuses on cell-based therapies or studies ATXN3 in unrelated disease contexts, indicating a significant gap in nutritional and peptide-based intervention research for this specific protein variant in Machado-Joseph disease.
Synthesis Agent (1)
Synthesis of 1 findings (peptides): The computational peptide design effort for ATXN3 wildtype Q27 (associated with Spinocerebellar atax...
Peptide Agent (1)
ATXN3 WILDTYPE: 1 candidate peptides designed