Rexis
  • Rexis: The L1 for DeSci
  • Key Challenges in BioDeSci
    • Overview
  • Data Level: Secure Biomedical Sharing
  • Model Level: Privacy-Preserving Fine-Tuning
  • Evaluation Level: Reproducible Computation
  • The Rexis Solution: Layer for DeSci
    • Overview
    • Data Level: Decentralized Biomedical Data Market
    • Model Level: Privacy-Preserving Training & Inference via Equivariant Encryption
    • Evaluation Level: Secure and Verifiable Biomedical Computation
  • Example BioDeSci Data and Models
    • Overview
  • Tabular Data
  • Biomedical Signals
  • Biological Sequences
  • Medical Imaging
  • Volumetric Medical Imaging
  • Spatial Omics Data
  • Tokenomics
    • $REX Overview
  • Links
    • rexis.io
  • Term of Use
  • Privacy Policy
  • Community
Powered by GitBook
On this page

Data Level: Secure Biomedical Sharing

Data Level: Secure and Efficient Biomedical Data Sharing

Biomedical research—especially in areas such as genomics, precision medicine, and clinical trials—critically depends on large-scale, diverse datasets from multiple institutions (Berger et al., 2019). Yet effective data sharing among hospitals, universities, and industry partners is often hampered by:

  • Heightened privacy risks

  • Strict regulatory requirements (e.g., HIPAA, GDPR)

  • Competitive and proprietary constraints (Nagaraj et al., 2020)

For example, genomic data can often be re-identified despite anonymization, creating serious privacy vulnerabilities that severely restrict data sharing and impede progress (Cho et al., 2022).

Common scenarios illustrate these barriers:

  • Multi-Institutional Clinical Studies Integrating patient records across institutions improves the robustness of clinical findings. However, privacy regulations frequently limit these exchanges, weakening collaboration and reproducibility (Berger et al., 2019).

  • Genomic Data Analysis Reluctance to share genomic datasets—due to re-identification risks—constrains large-scale studies essential for biomarker discovery and precision diagnostics (Cho et al., 2022).

  • Drug Discovery Collaborations Pharmaceutical companies and academic labs often withhold proprietary data (e.g., compound libraries, assay results) due to intellectual property concerns, slowing progress in translational research (Kim et al., 2021).

Several privacy-preserving computation methods have been proposed, but each exhibits key limitations when applied to biomedical data sharing:

  • Homomorphic Encryption (HE) allows computation on encrypted data but imposes substantial computational overhead—particularly for nonlinear models (Gentry, 2009; Gilad-Bachrach et al., 2016).

  • Differential Privacy (DP) protects individuals by injecting noise into results. However, this noise may degrade accuracy and diagnostic value (Abadi et al., 2016).

  • Secure Multi-party Computation (SMPC) avoids raw data sharing but suffers from high communication costs, making it less scalable for applications like genome-wide studies or multi-site trials (Cho et al., 2022).

These limitations highlight the need for more efficient, scalable, and privacy-preserving methods to enable biomedical data sharing at population scale.

References

  • Abadi, M., et al. (2016). Deep learning with differential privacy. CCS.

  • Berger, B., et al. (2019). Federated learning in biomedical research.

  • Cho, H., et al. (2022). Privacy-preserving genomic analysis. Annual Review of Biomedical Data Science.

  • Gentry, C. (2009). Fully homomorphic encryption using ideal lattices. STOC.

  • Gilad-Bachrach, R., et al. (2016). CryptoNets: Applying neural networks to encrypted data. ICML.

  • Kim, M., et al. (2021). Privacy-preserving federated learning in medicine. JAMIA.

  • Nagaraj, A., et al. (2020). Privacy constraints in clinical collaborations.

PreviousOverviewNextModel Level: Privacy-Preserving Fine-Tuning

Last updated 1 month ago