PyFasma

PyFasma integrates preprocessing tools, multivariate techniques, and spectral deconvolution within a modular framework for the reproducible and robust analysis of Raman spectroscopy data from biological samples.

Key Features:

• Preprocessing Tools: PyFasma includes spike removal, smoothing, baseline correction, and normalization for Raman spectral data.
• Multivariate Analysis: PyFasma implements PCA and PLS-DA for dimensionality reduction and classification of spectral datasets.
• Spectral Deconvolution: PyFasma provides methods for deconvolving complex Raman spectra of biological samples.
• Model Validation: PyFasma supports repeated stratified cross-validation to enhance the generalizability of multivariate analyses.

Scientific Applications:

• Biomedical Case Study: PyFasma distinguished healthy and osteoporotic cortical bone spectra, revealing statistically significant differences in mineral-to-matrix ratio and crystallinity between groups.

Methodology:

PyFasma employs preprocessing steps including spike removal, smoothing, baseline correction, and normalization, followed by multivariate analysis using PCA and PLS-DA, with model validation via repeated stratified cross-validation.