Probabilistic Data Structures for Secure Genomic Information Sharing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Private organizations are hesitant to share sensitive genomic information due to its sensitive nature, leading to a lack of valuable data for genomic analytics, as existing solutions either forego important information or obscure it to the point of being useless.
Innovation Solution
The use of probabilistic data structures, specifically Bloom filters, to encode and anonymize genomic information, allowing for secure sharing and analysis while maintaining a 0% false negative return rate and adjustable false positive rate, enabling rapid insertion and lookup with reduced storage requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If genomic information is shared freely, then the quantity of data for genomic analytics is improved, but the sensitivity and security risk of the information worsens
Solution Approach 1:
A probabilistic data structure acts as an intermediary between the genomic data and the analytics system. The structure encodes genomic sequences in a way that prevents direct reconstruction of the original sensitive information while still enabling analytical queries, thus mediating between data sharing and security concerns
Solution Approach 2:
Instead of sharing the original genomic data, the system creates and shares a probabilistic copy (data structure) that preserves analytical utility while removing the ability to reconstruct the exact original sequences. This copy contains encoded information sufficient for analytics but insufficient for identifying specific sensitive data
2Object-affected harmful factors
If genomic information is redacted or obscured, then the security of the information is improved, but the value for genomic data analytics deteriorates
Solution Approach 1:
The system transforms the genomic data by changing its parameter representation from raw sequences to probabilistic data structures. This transformation maintains the essential analytical properties (such as sequence similarity and composition) while altering the form to prevent direct interpretation of sensitive information
3Measurement precision
If traditional data structures are used for genomic data, then the accuracy of data representation is improved, but the storage requirements and processing time worsen
Solution Approach 1:
The genomic data is segmented into k-mers (short subsequences) which are then encoded in the probabilistic data structure. This segmentation allows the system to represent large genomic datasets in a compressed form while maintaining the ability to perform accurate analytical operations on the segmented units
Solution Approach 2:
The system changes the parameter representation from storing complete genomic sequences to storing probabilistic indicators of k-mer presence. This parameter transformation dramatically reduces storage requirements while preserving the ability to perform genomic analytics through probabilistic queries
Data Source
AI summary
Techniques for securely encoding, communicating, and comparing genomic information using probabilistic data structures are provided. In some embodiments, genomic information in a secure computing environment may be encoded and/or anonymized by building a probabilistic data structure that represents sub-strings of the genomic information as members of a set; the probabilistic data structure may then be securely transmitted outside the secure computing environment. In some embodiments, a probabilistic data structure representing sub-strings of sensitive genomic information as members of a set may be received in an unsecure computing environment and may be queried to generate output data indicating whether reference sub-strings are probable members of the set. In some embodiments, querying the probabilistic data structure, and other techniques of analyzing the probabilistic data structure, may be used to determine whether the sensitive genomic information corresponds to an organism associated with the reference genomic information.