FTIR Spectroscopy of Biofluids for Disease Diagnosis: Data Quality Influence on Group Separation

Fourier Transform Infrared (FTIR) Spectroscopy, particularly using the Attenuated Total Reflection (ATR) technique, has become a powerful tool in the analysis of biofluids for diagnostic purposes. The effectiveness of this technique in disease diagnosis relies heavily on the quality of the acquired spectral data. This study investigates the impact of key experimental parameters, such as the number of scans and spectral resolution, on the quality of FTIR spectra. By analyzing blood serum samples from different groups, we explore how these parameters influence the performance of Principal Component Analysis (PCA) and Support Vector Machine (SVM) classifiers in clustering tasks. Our findings indicate that while higher resolution does not necessarily enhance clustering, an increased number of scans can significantly improve classification performance when combined with lower-resolution settings. This work underscores the importance of optimizing FTIR data acquisition parameters to build robust predictive models in clinical diagnostics.