PAPEETE: Private, Authorized, and Fast Personal Genomic Testing

Angelo Massimo Perillo, Emiliano De Cristofaro

Abstract

As genomic sequencing becomes increasingly affordable, the demand for direct-to-consumer (DTC) genetic testing has surged. However, the sensitive nature of genomic data necessitates robust privacy protections. PAPEETE is a novel cryptographic protocol that enables personal genomic testing while simultaneously protecting user privacy, safeguarding a testing facility's intellectual property, and ensuring regulatory oversight. By leveraging additively homomorphic encryption, it allows for complex genetic calculations that are fast enough to run on a smartphone while being officially authorized by a third party.

The Challenge of Private Genomic Testing

Genomic data is uniquely identifying and reveals sensitive information about an individual's health predispositions and hereditary traits. Current privacy-preserving methods often struggle with three main issues:

  1. Scope: Many support only simple tests rather than the complex weighted averages required for modern medicine.
  2. Efficiency: Processing millions of Single Nucleotide Polymorphisms (SNPs) is often too slow for practical use.
  3. Authorization: There is a lack of mechanisms to ensure testing companies do not perform unauthorized "fishing expeditions" on sensitive genetic markers.

The PAPEETE Architecture

PAPEETE addresses these challenges through a three-party architecture involving a User, a Testing Facility, and a Certification Authority (CA) (e.g., the FDA).

Figure 1: The PAPEETE three-party architecture involving a User, Testing Facility, and Certification Authority.

The protocol utilizes Additively Homomorphic Elliptic Curve El-Gamal (AH-ECC), allowing additions on encrypted data without decryption. The process occurs in two phases:

  • Offline Authorization: The CA authorizes the weights used in the genetic test, ensuring the Testing Facility only runs approved tests.
  • Online Testing: Authorized, encrypted weights are sent to the User's device (e.g., a smartphone), which computes the weighted average against local genomic data. This ensures the User never sees the proprietary weights, and the Facility never sees the User's raw DNA.

Performance and Comparative Analysis

The authors evaluated PAPEETE using simulated genomic data ranging from 1,000 to 1,000,000 SNPs with a realistic 20% non-zero SNP rate. Prototype testing on a Google Nexus 5 smartphone demonstrated that the protocol is highly efficient. For a test involving 1 million SNPs, the online portion takes less than 19 minutes and requires only 63MB of data transfer.

Crucially, PAPEETE provides a unique combination of features compared to existing state-of-the-art methods.

Figure 5: Comparison of PAPEETE against existing works across key dimensions of privacy, authorization, efficiency, and support for weighted averages.

As shown in the comparison, PAPEETE is the first protocol to simultaneously provide privacy, authorization, efficiency, and support for weighted average calculations. The addition of the authorization layer introduces negligible overhead, increasing running time by only about 1% compared to non-authorized protocols.

Conclusion and Future Directions

PAPEETE demonstrates that robust privacy and regulatory authorization do not have to come at the cost of performance. Future work will focus on developing a user-friendly graphical interface, conducting human-factor studies to assess user trust, and expanding the protocol to support more exotic types of genomic analysis, such as structural variants.

Summary updated

BibTeX
@inproceedings{perillo2018papeete,
  author = {Angelo Massimo Perillo and Emiliano De Cristofaro},
  title = {{PAPEETE: Private, Authorized, and Fast Personal Genomic Testing}},
  booktitle = {ICETE},
  year = {2018},
  doi = {10.5220/0006859706500655},
}
Figure