BioDRUMs: An Open-Source Python Pipeline for Automated Mean Drug-to-Antibody Ratio and Biologics Structural Integrity Analysis from High-Resolution Mass Spectrometry Data

Abstract Drug-to-antibody ratio and structural integrity represent two critical quality attributes for antibody-drug conjugates and multispecific biologics in development. Intact mass spectrometry analysis of biologics is crucial to understanding the behavior of a new candidate, especially for complex modalities. Sample preparation is relatively robust and easy to reproduce in different laboratories and by different operators. However, mass spectrometric analysis and data processing represent the two most limiting steps. So far, no automated open-source pipeline has been published, which can automatically postprocess deconvoluted spectrum outputs from different vendors, plot data, and provide an overall interpretation of mean DAR and deep ADC characterization, especially on in vivo samples. BioDRUMs offers different features that complement existing commercial mass spectrometry deconvolution tools, such as customizable degradation-hypothesis generation, isobaric-species ambiguity reporting, vendor-neutral postprocessing of deconvoluted mass tables, structural-integrity scoring, automated plotting, and reporting. Streamlining data analysis of the mean drug-to-antibody ratio and structural integrity can accelerate and provide uniform data outputs and result delivery across different in vitro/in vivo studies. Moreover, a user-friendly pipeline would enable scientists to perform such an analysis for the first time more confidently, without requiring any pre-existing coding skills. In this work, we present BioDRUMs (Biologics Drug Ratio and Unified intact Mass analysis), an automated Python pipeline to simplify data analysis of the mean drug-to-antibody ratio and structural integrity analysis of biologics and bioconjugates. A graphical user interface (GUI) and a Web app interface have also been implemented to make it user-friendly for users with little to no prior programming knowledge.