ATR-FTIR Spectroscopy with Optimized Spectral Ranges Discriminates Stages of Cervical Cancer Progression


Description

Complete ATR-FTIR spectral dataset used in "ATR-FTIR Spectroscopy with Optimized Spectral Ranges Discriminates Stages of Cervical Cancer Progression." Contains 35 spectra (3736 wavenumbers, 399-4001 cm⁻¹) from healthy cervical cells (PC, n=12), primary cervical carcinoma (M1, n=12), and metastatic cells (M2, n=11). CSV format: first column = wavenumber, subsequent columns = sample spectra. Note: Samples M1_8 and PC_5 were identified as outliers and excluded from analysis (QC'd dataset: n=33, 11 per group).

Steps to reproduce

1. Load CSV data (wavenumber + 35 spectra) 2. Remove outliers M1_8 and PC_5 via PCA (>3 SD; verified by sensitivity testing) 3. Apply RMieS correction and vector normalization 4. Perform PCA-LDA classification with LOO cross-validation 5. Test different spectral ranges (900-1800, 900-1200, 700-2200 cm⁻¹) Software: Python 3.9+, scikit-learn, pandas, numpy, scipy

Institutions

• Ariel University

  Judea and Samaria Area, Ariel

Categories

Licence