Case Study
Transforming Rare Disease Treatment Decisions Through Curated Genomic Intelligence
A collaboration between Rancho BioSciences and Rady Children’s Institute for Genomic Medicine (RCIGM)
Download the PDF
Every year, up to 10% of the nearly 3.6 million newborns in the United States are hospitalized, and many require intensive care. About one third of these cases have an underlying genetic condition that may respond to targeted treatment. In these critical moments, clinicians need rapid access to trusted, evidence-based insights to guide treatment decisions.
Rady Children’s Institute for Genomic Medicine (RCIGM) developed the Genome-to-Treatment (GTRx) platform to deliver fast, accurate clinical guidance for rare genetic diseases. Rancho BioSciences partnered with RCIGM to build the scientific data foundation that powers it through comprehensive knowledge mining, evidence curation, and rigorous quality control. The result is a high-fidelity treatment knowledge base that supports real-time clinical decision making and sets a new standard for evidence readiness in genomic medicine.
The Challenge
Treatment evidence for rare diseases is fragmented across thousands of scientific articles, clinical trials, guidelines, and biochemical reports. Clinicians and researchers face the dual challenge of locating this information and judging its reliability for pediatric use. The complexity and unstructured nature of biomedical literature can slow the path from diagnosis to action.
To support rapid genome sequencing programs, newborn screening, and acute pediatric decision making, RCIGM required a scientifically rigorous methodology that could transform dispersed data into structured, actionable, and clinically ready intelligence.
The Rancho Solution
Rancho applied a proven, high-precision scientific data methodology designed for complex biomedical evidence. The approach combines deep domain expertise with structured governance, reproducible workflows, and multi-level review.
Comprehensive Knowledge Mining
- Systematic capture of treatment evidence across PubMed, conference abstracts, clinical trials, guidelines, and industry publications.
- Coverage of 1,040 gene-disease pairs with full traceability to source literature.
Expert Scientific Curation
- Evaluation of treatment intent, pediatric appropriateness, contraindications, precautions, and estimated time to treat.
- Classification of treatments as curative, effective or ameliorative, still in trials or unknown, or contraindicated.
- Direct alignment with target gene variant evidence.
Rigorous Quality, Governance, and Reviewer Alignment
- Independent scientific review of every curated treatment entry, with a governance team validating evidence integrity, data completeness, and consistency across genes and disorders.
- Structured workflows in REDCap to record edits, compare reviewer assessments, and finalize clinical recommendations.
- Annual review cycles that keep the data current with scientific knowledge.
Ready for Clinical Decision Support
- Rancho’s harmonized knowledge base integrates into the BeginNGS platform, where clinicians can access treatment insights within minutes of diagnosis.
Impact: From Hours to Minutes
The RCIGM and Rancho collaboration created a reliable, scalable infrastructure for rapid treatment decision making.
- disorders reviewed for ICU use
- 585
- disorders reviewed for newborn screening
- 455
- gene-disease pairs curated and available for clinical use
- 1,040
- treatments retained after expert review, out of 18,003 identified (11.3%)
- 2,043
- positive cases in the sequencing program benefited directly from curated treatment options
- 28 of 28
- of more than 750 children in an ongoing Phase 3 trial diagnosed with rare diseases supported by curated insights
- 3.7%
Curation time per gene-disease pair dropped from 15 hours in 2020 to under 6 hours in 2025.
The Downstream Value of High-Quality Curated Evidence
- Clinicians receive treatment guidance in minutes, not hours.
- Newborns gain faster access to evidence-based interventions.
- Evidence quality and consistency improve across clinical teams.
- The dataset supports expansion into large-scale newborn screening programs.
- Rancho created a reusable evidence framework valuable across many therapeutic areas.
- The methodology is directly applicable to biopharma use cases in real-world evidence, translational research, safety, and AI model training.
Scaling Treatment Intelligence Across Leading Health Systems
Multiple institutions across the GTRx and BeginNGS networks, including Balboa Naval Medical Center, Grossmont Hospital, Loma Linda Medical Center, Rancho Springs Medical Center, Scripps Medical Network, and UCSD Medical Network, rely on this curated dataset to support clinical decision making.
A Blueprint for Biopharma Evidence Generation and AI Readiness
Biopharma organizations face similar challenges in synthesizing biomedical literature, clinical evidence, and mechanistic knowledge across therapeutic areas. Rancho’s work with RCIGM demonstrates:
- The ability to convert unstructured data into high-quality scientific intelligence.
- A scalable methodology that supports precision medicine and rare disease programs.
- A foundation for training AI, LLMs, and agentic decision systems with trusted data.
- A repeatable framework for evidence generation across discovery, translational research, and clinical development.
Rancho delivers the scientific depth and data discipline required to accelerate insights, reduce uncertainty, and support more confident, data-driven decisions across the R&D lifecycle.
Conclusion
By combining genomic expertise, scientific curation, and rigorous data engineering, Rancho BioSciences enabled RCIGM to transform the speed and quality of rare disease treatment decisions. This collaboration shows how high-integrity data and expert curation can unlock meaningful clinical impact, and serves as a model for biopharma organizations seeking to scale evidence development, improve decision quality, and prepare their R&D ecosystems for the next generation of AI-powered scientific discovery.
%20(1).png)