MAGPIE: A Machine Learning Approach to Decipher Protein–Protein Interactions in Human Plasma

• DOI: 10.1021/acs.jproteome.4c00160
• Publication Date: 2025-01-07T16:42:21Z
• Authors: Emily Hashimoto-Roth; Diane Forget; Vanessa P. Gaspar; Steffany A. L. Be...; Marie-Soleil Gaut...; Benoit Coulombe; Mathieu Lavallée-...
• Keywords: Immunoprecipitation; DECIPHER

Immunoprecipitation coupled to tandem mass spectrometry (IP-MS/MS) methods are often used identify protein–protein interactions (PPIs). While these approaches prone false positive identifications through contamination and antibody nonspecific binding, their results can be filtered using negative controls computational modeling. However, such filtering does not effectively detect false-positive when IP-MS/MS is performed on human plasma samples. Therein, proteins cannot overexpressed or inhibited, existing modeling algorithms adapted for execution without controls. Hence, we introduce MAGPIE, a novel machine learning-based approach identifying PPIs in IP-MS/MS, which leverages that include antibodies targeting expected present plasma. A set of interaction first constructed. MAGPIE then assesses the reliability detected experiments target known proteins. When applied five as proof concept, our algorithm identified 68 with an FDR 20.77%. significantly outperformed state-of-the-art PPI discovery tool predicted PPIs. Our provides unprecedented ability PPIs, enables better understanding biological processes