Author Correction: Truly privacy-preserving federated analytics for precision medicine with multiparty homomorphic encryption

David Froelicher, Juan Ramón Troncoso-Pastoriza, Jean Louis Raisaro, Michel A. Cuendet, João Sá Sousa, Hyunghoon Cho, Bonnie Berger, Jacques Fellay, Jean‐Pierre HubauxNature Communications 2021 · DOI · OpenAlex

Overview

This document serves as a formal correction to the research article "Truly privacy-preserving federated analytics for precision medicine with multiparty homomorphic encryption," originally published in Nature Communications. The primary purpose of this correction is to ensure proper academic attribution and to provide a more comprehensive overview of the existing literature in the field of secure genomic data analysis.

The Nature of the Correction

The authors have identified an omission in the original publication's literature review. Specifically, a significant prior work by Blatt et al. (2020), titled "Secure large-scale genome-wide association studies using homomorphic encryption," was not cited in the initial version of the paper.

This missing reference is critical because it represents an existing "sophisticated solution" for Federated Analytics (FA). In the context of precision medicine, FA requires end-to-end privacy protection that covers not only the raw input data but also the "shared intermediate data" generated during the analysis process.

Contextualizing Federated Analytics in Precision Medicine

The original research focuses on the development of "truly privacy-preserving" federated analytics. In the domain of precision medicine, FA is essential for analyzing sensitive patient data across multiple institutions without compromising individual privacy.

Genome-wide association studies (GWAS) are a cornerstone of this field, helping researchers identify genetic variants associated with specific diseases. However, because genomic data is uniquely identifying, sharing it across institutions poses significant privacy risks. Federated Analytics addresses this by allowing data to remain local while only sharing insights or intermediate computations.

The Role of Multiparty Homomorphic Encryption

Multiparty Homomorphic Encryption (MHE) is a specialized cryptographic technique that allows multiple parties to jointly compute functions over their collective data while keeping the data itself encrypted. This is particularly relevant for precision medicine, where collaborative research is necessary but data privacy is paramount. By integrating MHE into federated analytics, researchers aim to create a system where no single entity ever sees the raw data of others, and even the intermediate steps of the computation remain shielded from prying eyes.

The Importance of End-to-End Privacy

A key challenge in FA is ensuring end-to-end privacy protection. This involves securing not only the raw input data at rest but also the "shared intermediate data" produced during the analysis process. If intermediate results are leaked, they can sometimes be used to reconstruct sensitive information about the original datasets.

The correction clarifies the state of the art by acknowledging that end-to-end privacy protection for intermediate data is a known challenge with existing proposed solutions. The work by Blatt et al. (2020) is a foundational step in addressing this within the context of GWAS using homomorphic encryption.

Academic Integrity and Field Evolution

The primary insight of this correction notice is the importance of citing foundational work to provide an accurate roadmap of how the field has evolved. By acknowledging prior sophisticated solutions, the authors place their own contributions—utilizing multiparty homomorphic encryption—within the broader context of ongoing efforts to achieve robust privacy in medical data analytics.

The original article (both PDF and HTML versions) has been updated to reflect this change, ensuring that the scholarly record is complete and that the contributions of Blatt et al. are properly recognized.

This summary is based on the official Author Correction notice for the paper published in Nature Communications.

Summary updated

BibTeX
@inproceedings{froelicher2021author,
  author = {David Froelicher and Juan Ramón Troncoso-Pastoriza and Jean Louis Raisaro and Michel A. Cuendet and João Sá Sousa and Hyunghoon Cho and Bonnie Berger and Jacques Fellay and Jean‐Pierre Hubaux},
  title = {{Author Correction: Truly privacy-preserving federated analytics for precision medicine with multiparty homomorphic encryption}},
  booktitle = {Nature Communications},
  year = {2021},
  doi = {10.1038/s41467-021-26885-6},
}

Loading related content...

Figure