Search Results
Correcting systematic errors by hybrid 2D correlation loss functions in nonlinear inverse modelling
2023, Mayerhöfer, Thomas G., Noda, Isao, Pahlow, Susanne, Heintzmann, Rainer, Popp, Jürgen
Recently a new family of loss functions called smart error sums has been suggested. These loss functions account for correlations within experimental data and force modeled data to obey these correlations. As a result, multiplicative systematic errors of experimental data can be revealed and corrected. The smart error sums are based on 2D correlation analysis which is a comparably recent methodology for analyzing spectroscopic data that has found broad application. In this contribution we mathematically generalize and break down this methodology and the smart error sums to uncover the mathematic roots and simplify it to craft a general tool beyond spectroscopic modelling. This reduction also allows a simplified discussion about limits and prospects of this new method including one of its potential future uses as a sophisticated loss function in deep learning. To support its deployment, the work includes computer code to allow reproduction of the basic results.
Spatially resolved spectroscopic differentiation of hydrophilic and hydrophobic domains on individual insulin amyloid fibrils
2016, Deckert-Gaudig, Tanja, Kurouski, Dmitry, Hedegaard, Martin A. B., Singh, Pushkar, Lednev, Igor K., Deckert, Volker
The formation of insoluble β-sheet-rich protein structures known as amyloid fibrils is associated with numerous neurodegenerative diseases, such as Alzheimer’s and Parkinson’s disease. A detailed understanding of the molecular structure of the fibril surface is of interest as the first contact with the physiological environment in vivo and plays a decisive role in biological activity and associated toxicity. Recent studies reveal that the inherent sensitivity and specificity of tip-enhanced Raman scattering (TERS) renders this technique a compelling method for fibril surface analysis at the single-particle level. Here, the reproducibility of TERS is demonstrated, indicating its relevance for detecting molecular variations. Consequently, individual fibrils are systematically investigated at nanometer spatial resolution. Spectral parameters were obtained by band-fitting, particularly focusing on the identification of the secondary structure via the amide III band and the differentiation of hydrophobic and hydrophilic domains on the surface. In addition multivariate data analysis, specifically the N-FINDR procedure, was employed to generate structure-specific maps. The ability of TERS to localize specific structural domains on fibril surfaces shows promise to the development of new fibril dissection strategies and can be generally applied to any (bio)chemical surface when structural variations at the nanometer level are of interest.