Unsupervised self-organising map of prostate cell Raman spectra shows disease-state subclustering

TL;DR

An unsupervised, self-organising map approach is used to analyse live-cell Raman spectroscopy data obtained from prostate cell-lines to test the feasibility of this method to differentiate, at the single-cell-level, cancer from normal using high-dimensional datasets with minimal preprocessing.

Abstract

Prostate cancer is a disease which poses an interesting clinical question: should it be treated? A small subset of prostate cancers are aggressive and require removal and treatment to prevent metastatic spread. However, conventional diagnostics remain challenged to risk-stratify such patients, hence, new methods of approach to biomolecularly subclassify the disease are needed. Here we use an unsupervised, self-organising map approach to analyse live-cell Raman spectroscopy data obtained from prostate cell-lines; our aim is to test the feasibility of this method to differentiate, at the single-cell-level, cancer from normal using high-dimensional datasets with minimal preprocessing. The results demonstrate not only successful separation of normal prostate and cancer cells, but also a new subclustering of the prostate cancer cell-line into two groups. Initial analysis of the spectra from each of the cancer subclusters demonstrates a differential expression of lipids, which, against the normal control, may be linked to disease-related changes in cellular signalling.

Figures (6)

• Figure 1: Raman spectra generated from thirty cell nuclei: fifteen from the normal prostate cell-line (PNT2-C2), and fifteen from the cancerous prostate cell-line (LNCaP). The spectral intensity has been normalised and is displayed in arbitrary units.
• Figure 2: SOM trained on the small Raman spectra dataset. A stripe of low density nodes across the central map separates the bulk of PNT2-C2 and LNCaP data.
• Figure 3: SOM built with the full Raman spectra dataset. There is good spread of data across the entire map, with low density stripes from the bottom left to top right, and from the centre left to the centre bottom.
• Figure 4: SOM from Figure \ref{['main_som']} with observations mapping to each node overlaid. The PNT2-C2 observations are located to the top left, separated from the LNCaP observations by a diagonal, low-nodal density stripe. A vertical low-density stripe in the centre of the map divides the LNCaP group into two subclusters.
• Figure 5: Mean Raman spectra of the full dataset and each cluster.