The Impact of Preprocessing Methods for a Successful Prostate Cell Lines Discrimination Using Partial Least Squares Regression and Discriminant Analysis Based on Fourier Transform Infrared Imaging

Abstract

Fourier transform infrared spectroscopy (FT-IR) is widely used in the analysis of the chemical composition of biological materials and has the potential to reveal new aspects of the molecular basis of diseases, including different types of cancer. The potential of FT-IR in cancer research lies in its capability of monitoring the biochemical status of cells, which undergo malignant transformation and further examination of spectral features that differentiate normal and cancerous ones using proper mathematical approaches. Such examination can be performed with the use of chemometric tools, such as partial least squares discriminant analysis (PLS-DA) classification and partial least squares regression (PLSR), and proper application of preprocessing methods and their correct sequence is crucial for success. Here, we performed a comparison of several state-of-the-art methods commonly used in infrared biospectroscopy (denoising, baseline correction, and normalization) with the addition of methods not previously used in infrared biospectroscopy classification problems: Mie extinction extended multiplicative signal correction, Eiler’s smoothing, and probabilistic quotient normalization. We compared all of these approaches and their effect on the data structure, classification, and regression capability on experimental FT-IR spectra collected from five different prostate normal and cancerous cell lines. Additionally, we tested the influence of added spectral noise. Overall, we concluded that in the case of the data analyzed here, the biggest impact on data structure and performance of PLS-DA and PLSR was caused by the baseline correction; therefore, much attention should be given, especially to this step of data preprocessing.

ATR-FTIR spectroscopy combined with multivariate analysis as a rapid tool to infer the biochemical composition of Ulva laetevirens (Chlorophyta)

“…As background, a sample-free spectrum was collected before each sample was recorded and then subtracted automatically from the sample spectra prior to further analysis. Every sample was recorded in triplicate using three new aliquots of powder, resulting in an average spectra from these triplicates (Oberreuter et al, 2002;Piermaria et al, 2011;Lee et al, 2017;Innes et al, 2019;Liberda et al, 2021). Between every measurement the ATR crystal plate was cleaned with ethanol (96%) and dried to ensure a clean crystal surface (Goḿez-Ordoñez and Rupeŕez, 2011).…”

“…Baseline correction provided for ATR in the Perkin Elmer Spectrum IR software version 10.6.1 was used. Multiplicative signal correction, present in "pls" 2.8-0 package (Mevik and Wehrens, 2007), is used as normalization method, as is also the method of choice in the literature for ATR-FTIR (Lee et al, 2017;Innes et al, 2019;Liberda et al, 2021). The spectral range of 2750-1800 cm −1 is excluded because of the low signal-tonoise ratio in this spectral region caused by the absorbance of the ATR crystal.…”

Ten quick tips for biomarker discovery and validation analyses using machine learning

“…Many artifacts and normalization issues are data type specific and need to be addressed using dedicated preprocessing and filtering methods. Tailored software solutions have been made available to preprocess clinical data [21], NGS data [49], microarray data [50], different types of metabolomics and proteomics data [18], and cellular and brain imaging data [51][52][53][54]. Although no generic rules and methods exist for all data types, the following considerations apply to most datasets.…”