Loureiro, Joe, Jennings, Lori, Aspesi Jr, PJ, Mapa, Felipa, Perry, Darryl, Parker, Jefferson, Williams, Blake, Su, Wendy, Tabacman, Eduardo, Delisle, Kirk, Avishan, Kayvan, Williams, Alan, Myers, Vic and Janjic, Nebojsa (2024) Modeling Proteomes Using Unsupervised Machine Learning Approaches on SomaScan Aptamer-based Proteome Hybridization Technology. bioRxiv.

Abstract

Abstract: Robust and reliable proteome measurements provide mechanistic insights important to biomedical research. We have developed a clinical proteome profiling platform named So-maScan that has expanded to measure 7,523 proteoforms for 6,594 human proteins by UniprotID. SomaScan is based on modified DNA-based affinity reagents called SOMAmers (Slow Off-rate Modified Aptamers). Providing independent corroborating evidence that the SomaScan assay measures the intended proteins enhances the utility and interpretability of the platform. We have therefore evaluated the capabilities of the platform by profiling a panel of well characterized CCLE cancer models. Unsupervised learning methods demonstrate the SomaScan assay’s ability to distinguish cell lines and identify both tissue-specific and oncogenic pathways. Publicly avail-able CCLE transcriptome profiling sets an expectation of endogenous protein to which each SOMAmer is annotated. We found orthogonal transcript data support protein annotation for ap-proximately one third of the SOMAmer reagents, consistent with transcript to protein correlation observed in many other studies. The SomaScan platform is a technically reproducible proteome measuring device available for biomedical and clinical applications providing investigators with reliable data to inform underlying biochemical mechanisms and with the ability to utilize individual SOMAmer reagents for follow-up studies and clinical biomarker assay development.

Item Type:	Article
Keywords:	SOMAmer, SomaScan, protein detection, clinical proteomics
Date Deposited:	20 Mar 2024 00:45
Last Modified:	20 Mar 2024 00:45
URI:	https://oak.novartis.com/id/eprint/51142