DeCure's autonomous Rare AI scientist is researching a drug-repurposing hypothesis for Stargardt disease — screening already-approved drugs against its 25-gene Open Targets disease module to publish open-access research. Research is fast; the path to publication is funded in milestone stages.
Disease moduleStargardt disease maps to a 25-gene Open Targets module — the target space DeCure's AI scientist screens approved drugs against.
DeCure.ai methodSignature reversal (LINCS) plus network proximity (STRING) rank already-approved drugs likely to perturb this module — the same engine that produces DeCure.ai's repurposing hypotheses.
Repurposing thesisScreening approved medicines against this disease module, then publishing the evidence for the strongest candidate. Known pharmacology and human exposure data make the first question sharper — they do not establish safety or efficacy in a new indication.
Research record
01
ResearchComing soon
Candidate research + dossier — target rationale, drug-repurposing thesis and evidence pack.proof: Published dossier + on-chain hash
02
ValidationComing soon
In-vitro biological validation at a contract research org (CRO).proof: CRO contract + in-vitro report
03
Peer review & paperComing soon
Peer-reviewed paper published open-access (preprint + journal).proof: DOI + open-access link + on-chain hash
Current lead
No approved-drug candidate for stargardt disease is corroborated in the literature DeepSearch retrieved. Some conditions are managed with non-pharmacological care — a device, surgery or physical therapy — rather than a medicine; that may be the case here, or the literature we found may simply be too sparse yet to support a drug-repurposing angle.
Molecular view
collagen type II alpha 1 chain (COL2A1) — COL2A1 is one of the genes genetically linked to this disease in Open Targets — shown as context, not as a drug target we're pursuing: no approved-drug candidate for this disease is yet corroborated in the literature we found.
Loading structure…
helix sheet p33drag to rotate · scroll to zoom
RCSB Protein Data Bank · entry 5NIR · 1.74 Å · ligand 3,6,9,12,15,18-HEXAOXAICOSANE-1,20-DIOL (P33). Experimental structure, not a prediction.
What the evidence adds up to
A 2020 study in Abca4−/− mice tested an overlapping AAV dual vector system for ABCA4 gene therapy. Subretinal injection of the dual vector mix, the separate 5′ and 3′ vectors, a GFP reporter vector, or diluent alone produced no structural or functional changes beyond those seen in the sham control cohort up to six months after injection. Histologic changes were limited to the superior retina at the injection site. The authors concluded the dual vector system showed no additional signs of toxicity beyond the sham injection and encouraged its future use in clinical trials. No human data were reported.
A 2010 study used spectral-domain OCT to measure peripapillary retinal nerve fibre layer thickness in 52 eyes of 27 Stargardt patients (mean age 38.3 years). Fourteen patients (51.9%) showed thinning in one or more quadrants in at least one eye, and four patients (14.8%) in both eyes. Five patients (18.5%) showed thickening in at least one eye, and four (14.8%) in both eyes. The authors recommended that Stargardt patients considered for treatment options have RNFL thickness measurements, but no treatment outcomes were assessed.
A 2025 meeting report from the Stargardt's Connected Research Network summarised progress in diagnosis, management, and treatments. Topics included natural history, genetic basis, animal models, emerging therapies, disease impact, technological advancements, lifestyle modifications, and stakeholder engagement. The network called for improved genetic testing and counselling, research registries, clinical trials, multidisciplinary care, lifestyle and dietary modifications, and technological support. No specific drug or therapy results were presented.
A second 2025 study developed a quantitative fundus autofluorescence score for paediatric STGD1 patients using 73 images from 14 individuals over up to six years and 27 control images from eight individuals over up to five years. The score showed 91% Spearman correlation with absolute age and 97% with estimated time from onset when averaged over both eyes. The authors noted that complete automation of the analysis remains difficult due to challenges in clinical-care images. The tool is intended for tracking natural progression and potentially the effects of genotype, environment, and therapeutic intervention, but no intervention was tested. What is still missing is a completed first-in-human trial of the dual vector system, validated biomarkers that can be used as surrogate endpoints in trials, and a trial design that accounts for the slow and variable progression of Stargardt disease, particularly in paediatric patients where stratification by genotype and disease stage is not yet standardised.
Evidence
Retrieved by DeepSearch across 234,678,978 indexed works and resolved on OpenAlex — ranked by citations, including the results that did not work.
JAMA Ophthalmology · 2017 · 108 citations · open access
Progression of Stargardt Disease as Determined by Fundus Autofluorescence in the Retrospective Progression of Stargardt Disease Study (ProgStar Report No. 9)
AbstractImportance: Sensitive outcome measures for disease progression are needed for treatment trials of Stargardt disease. Objective: To describe the yearly progression rate of atrophic lesions in the retrospective Progression of Stargardt Disease study. Design, Setting, and Participants: A multicenter retrospective cohort study was conducted at tertiary referral centers in the United States and Europe. A total of 251 patients aged 6 years or older at baseline, harboring disease-causing variants in ABCA4 (OMIM 601691), enrolled in the study from 9 centers between August 2, 2013, and December 12, 2014; of these patients, 215 had at least 2 gradable fundus autofluorescence images with atrophic lesion(s) present in at least 1 eye. Exposures: Areas of definitely decreased autofluorescence (DDAF) and questionably decreased autofluorescence were quantified by a reading center. Progression rates were estimated from linear mixed models with time as the independent variable. Main Outcomes and Measures: Yearly rate of progression using the growth of atrophic lesions measured by fundus autofluorescence. Results: A total of 251 participants (458 study eyes) were enrolled. Images from 386 eyes of 215 participants (126 females and 89 males; mean [SD] age, 29.9 [14.7] years; mean [SD] age of onset of symptoms, 21.9 [13.3] years) showed atrophic lesions present on at least 2 visits and were graded for 2 (156 eyes), 3 (174 eyes), or 4 (57 eyes) visits. A subset of 224 eyes (123 female participants and 101 male participants; mean [SD] age, 33.0 [15.1] years) had areas of DDAF present on at least 2 visits; these eyes were included in the estimation of the progression of the area of DDAF. At the first visit, DDAF was present in 224 eyes (58.0%), with a mean (SD) lesion size of 2.2 (2.7) mm2. The total mean (SD) area of decreased autofluorescence (DDAF and questionably decreased autofluorescence) at first visit was 2.6 (2.8) mm2. Mean progression of DDAF was 0.51 mm2/y (95% CI, 0.42-0.61 mm2/y), and of total decreased fundus autofluorescence was 0.35 mm2/y (95% CI, 0.28-0.43 mm2/y). Rates of progression depended on the initial size of the lesion. Conclusions and Relevance: In Stargardt disease with DDAF lesions, fundus autofluorescence may serve as a monitoring tool for interventional clinical trials that aim to slow disease progression. Rates of progression depended mainly on initial lesion size.
The Natural History of Stargardt Disease with Specific Sequence Mutation in the<i>ABCA4</i>Gene
AbstractPURPOSE: To determine longitudinal changes in fundus appearance and visual function in patients with Stargardt with at least one allelic mutation (Gly1961Glu) in the ABCA4 gene. METHODS: Sixteen patients with a diagnosis of Stargardt disease and a Gly1961Glu mutation were enrolled. All patients underwent a complete ocular examination including best corrected visual acuity, Goldmann visual field (GVF), and full-field ERG examinations. The percentage of patients who showed at least a doubling in the log of the minimum angle of visual resolution (logMAR) between their initial and most recent visits was determined, as was the percentage of patients who showed a doubling in the size of the central scotoma over this duration. RESULTS: Nine patients had at least a doubling of the logMAR visual acuity in their right eyes and 10 patients in their left eyes, over a mean follow-up (FU) period of 18.6 years. Of 15 patients, 46.7% had equal to or more than a doubling of the central scotoma area in response to a II2e test stimulus in the right eye and 60.0% in the left eyes. Among 10 patients whose ERGs were initially normal for rod and cone responses, 8 remained normal at their most recent FU visit. CONCLUSIONS: In these patients with Stargardt disease and a Gly1961Glu mutation, most showed a clinical phenotype characterized by fundus changes localized to the foveal and parafoveal regions, normal ERG amplitudes, absence of a silent or masked choroid, and a mean age at initial presentation in the third decade.
Assessment of AAV Dual Vector Safety in the <i>Abca4<sup>−/−</sup></i> Mouse Model of Stargardt Disease
AbstractPurpose: Adeno-associated viral (AAV) gene therapy treatment for Stargardt disease currently requires a dual vector approach owing to the size of the ATP-binding cassette transporter family member gene (ABCA4). The nature of the dual vector system creates the potential for adverse events. Here we have investigated an overlapping adeno-associated viral ABCA4 dual vector system for signs of toxicity in Abca4−/− mice as a prelude to dual vector first in human clinical trials. Methods: Abca4−/− mice received a subretinal injection of a 1:1 5′:3′ dual vector mix; 5′ vector only; 3′ vector only; a GFP reporter vector; or diluent only (sham). All vectors were adeno-associated virus-8 Y733F. Mice were subsequently assessed for signs of toxicity as measured by loss in retinal structure by optical coherence tomography and retinal function by electroretinography up to 6 months after injection. Results: Subretinal delivery of the dual vector system and its comprising parts induced no structural or functional changes relative to paired uninjected eyes beyond those observed in the sham control cohort. Histologic changes were limited to the superior retina where the injection was performed. Electroretinography analysis confirmed the dual vector system inferred no functional changes beyond those observed in the sham control cohort. Conclusions: An optimized overlapping dual vector system for the treatment of Stargardt disease shows no additional signs of toxicity beyond those observed from a sham injection. Translational Relevance: This presentation of safety of a dual vector system for the treatment of Stargardt disease encourages its future use in clinical trial.
British Journal of Ophthalmology · 2010 · 17 citations · open access
Spectral-domain OCT peripapillary retinal nerve fibre layer thickness measurements in patients with stargardt disease
AbstractAIMS: To evaluate the presence of peripapillary retinal nerve fibre layer (RNFL) defects in patients with Stargardt disease by using spectral-domain optical coherence tomography (SD-OCT). METHODS: Fifty-two eyes of 27 patients with Stargardt disease underwent peripapillary RNFL thickness measurements using SD-OCT. RESULTS: Twenty-seven patients with Stargardt disease were enrolled. Their mean (±SD) age was 38.3 (14.7) years. Fourteen patients (51.9%) showed a thinning of the peripapillary RNFL in one or more quadrants in at least one eye, and four patients (14.8%) in both eyes. Five patients (18.5%) showed a thickening of the peripapillary RNFL in at least one eye, and four patients (14.8%) in both eyes. CONCLUSION: This study demonstrated the presence of defects in the peripapillary RNFL thickness in patients with Stargardt disease by using SD-OCT. It would be clinically prudent that Stargardt patients considered for various treatment options be considered for RNFL thickness measurements.
Stargardt's Connected Research Network Inaugural Meeting: Landscape Review and Horizon Scanning of Stargardt Disease
AbstractPurpose: The purpose of this study was to update the recent progress in the diagnosis, management, and treatments for Stargardt disease. Methods: On November 22, 2024, Stargardt's Connected held its inaugural meeting of their Research Network, attended by clinicians, researchers, industry partners, and patient representatives. This meeting aimed to provide an update on Stargardt disease management and research and develop a call to action for the wider community. The format was rapid-fire presentations followed by a roundtable discussion. This review presents the meeting proceeding, along with a summary of best up-to-date evidence and key calls to action. Results: Topics included: (1) advances in the understanding of Stargardt disease: natural history, genetic basis, animal models, and emerging therapies; (2) supporting individuals with Stargardt disease: disease impact, technological advancements, and lifestyle modifications; and (3) advancing research through stakeholder engagement, research registries, and patient input. The network acknowledges the importance of collaboration among patients, clinicians, researchers, and industry to address critical gaps in the diagnosis, management, and treatment of Stargardt disease. Patient engagement was emphasized as being crucial for driving progress in the field. Conclusions: Recent years have seen significant progress in understanding and managing Stargardt disease. Key calls to action included the areas of improving genetic testing and counseling, advancing research registries, supporting research and clinical trials, growing multidisciplinary care, addressing lifestyle and dietary modifications, and enhancing technological support. Translational Relevance: This Stargardt's Connected Research Network initiative outlines a multistakeholder engagement model to discuss ongoing research and emerging treatments for Stargardt disease, a condition with promising therapeutic developments.
Quantifying the Progression of Stargardt Disease in Double-Null ABCA4 Carriers Using Fundus Autofluorescence Imaging
AbstractPurpose: To score real-world fundus autofluorescence (FAF) images of pediatric patients with ABCA4-related Stargardt disease (STGD1), in a way that is automatable, scales with the disease progression, and is applicable to a wide time interval in the natural history of the disease. Methods: We developed the score based on a series of Optos wide-field FAF images of pediatric STGD1 patients (73 images; 14 individuals) and controls (27 images; 8 individuals). The patients' images were obtained over up to 6 years, and the controls over up to 5 years. In each image, we manually selected an artifact-free region, within which we evaluated an average of the pixel-level intensity score, constructed so that the average increases with progression of the disease. Results: The score we propose provides a statistically robust measure of disease progression (91% Spearman correlation with the absolute age, 97% with the estimated time from onset, when averaged over both eyes), comparable across timepoints and patients. Conclusions: FAF is a reliable tool in STGD1 diagnostics, but its quantitative description must be modified to be applicable to tracking the disease progression. Analyzing images obtained in the course of clinical care of pediatric patients poses special challenges that make complete automation difficult. Translational Relevance: Our methodology provides a quantitative tool for investigating the natural progression of the Stargardt disease, and, potentially, the effects of genotype, environment, and therapeutic intervention on its course.
Disease module: DeepOracle (Open Targets). Structures: RDKit from PubChem SMILES. Literature: retrieved by DeepSearch across 234,678,978 indexed works (targeted per-candidate search), resolved on OpenAlex.
DeCure is a research and publication project, not medical advice and not a treatment. "DeCure for X" describes a research goal, not a claim that a cure exists. Backing a cure is a contribution to fund the research — it is not an investment, and confers no yield, royalty, equity or IP ownership. Papers are published open-access by the DeCure.ai DAO.
