Parallel and private generalized suffix tree construction and query on genomic data.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomic Data Analysis

Background:

• Digitization of healthcare data yields large genomic datasets, advancing disease understanding.
• Genomic data presents storage and retrieval challenges, with privacy concerns limiting access and discovery.

Purpose of the Study:

• To propose a privacy-preserving string query execution framework for large-scale genomic data.
• To enhance secure indexing and querying of encrypted genomic data outsourced to cloud servers.

Main Methods:

• Utilized Generalized Suffix Trees (GSTs) for efficient, scalable construction on large genomic datasets.
• Developed a privacy-preserving framework with a tree-based hashing mechanism for secure string operations.
• Enabled secure execution of exact and set-maximal string matches on encrypted genomic data.

Main Results:

• Experimental results demonstrate the scalability of the proposed GST-based framework in a cloud environment.
• The method achieved a 4x speedup compared to the state-of-the-art Burrows-Wheeler Transformation (BWT) method.
• Set-maximal matches were executed in approximately 36.7s, significantly faster than the BWT method's 160.85s.

Conclusions:

• The proposed privacy-preserving framework offers a scalable and efficient solution for querying large genomic datasets.
• This approach addresses privacy concerns while facilitating timely scientific discovery in genomics.