PATHE: A Privacy-Preserving Mass Spectrometry Database Pattern Search Platform with Fully Homomorphic Encryption

Many companies maintain extensive, high-quality databases to deliver services, making preserving data privacy during database searches crucial. For instance, pharmaceutical companies heavily depend on database search methods to identify and analyze unknown compounds by comparing them with the reference datasets. Fully Homomorphic Encryption (FHE) provides a delicate balance between companies' data privacy and the public's database pattern search. With FHE, the server can take encrypted queries from clients and search through the reference database without decryption, thus guaranteeing data security. In this work, we propose PATHE which exploits FHE and hyperdimensional computing (HDC) for high-performance privacy-preserving database search. We test PATHE on a large-scale proteomic open modification search (OMS) in an encrypted mass spectrometry library. PATHE takes advantage of the simplicity of HDC, which consists of basic vector operations without complex arithmetic, to create an FHE-friendly design with shallow data dependencies. To our knowledge, PATHE is the first to combine FHE with HDC-based database search and achieves at least 2201.7× speedup compared to prior FHE search algorithm.