Rapid evaluation of immobilized immunoglobulins using automated mass-segmented ToF-SIMS

Surface interactions largely control how biomaterials interact with biology and how many other types of materials function in industrial applications. ToF-SIMS analysis is extremely useful for interrogating the surfaces of complex materials and shows great promise in analyzing biological samples. Previously, the authors demonstrated that segmentation (between 1 and 0.005 m/z mass bins) of the mass spectral axis can be used to differentiate between polymeric materials with both very similar and dissimilar molecular compositions. Here, the same approach is applied for the analysis of proteins on surfaces, focusing on the effect of binding and orientation of an antibody on the resulting ToF-SIMS spectrum. Due to the complex nature of the samples that contain combinations of only 20 amino acids differing in sequence, it is enormously challenging and prohibitively time-consuming to distinguish the minute variances presented in each dataset through manual analysis alone. Herein, the authors describe how to apply the newly developed rapid data analysis workflow to previously published ToF-SIMS data for complex biological materials, immobilized antibodies. This automated method reduced the analysis time by two orders of magnitudes while enhancing data quality and allows the removal of any user bias. The authors used mass segmentation at 0.005 m/z over a 1-300 mass range to generate 60 000 variables. In contrast to the previous manual binning approach, this method captures the entire mass range of the spectrum resulting in an information-rich dataset rather than specifically selected mass spectral peaks. This work constitutes an additional proof of concept that rapid and automated data analyses involving mass-segmented ToF-SIMS spectra can efficiently and robustly analyze a broader range of complex materials, ranging from generic polymers to complicated biological samples. This automated analysis method is also ideally positioned to provide data to train machine learning models of surface-property relationships that can greatly enhance the understanding of how the surface interacts with biology and provides more accurate and robust quantitative predictions of the biological properties of new materials.