Raman spectroscopy as a tool for ecology and evolution

Abstract

Scientists are always on the lookout for new modalities of information which could reveal new biological features that are useful for deciphering the complexity of biological systems. Here, we introduce Raman spectroscopy as a prime candidate for ecology and evolution. To encourage the integration of this microscopy technique in the field of ecology and evolution, it is crucial to discuss first how Raman spectroscopy fits within the conceptual, technical and pragmatic considerations of ecology and evolution. In this paper, we show that the spectral information holds reliable indicators of intra- and interspecies variations, which can be related to the environment, selective pressures and fitness. Moreover, we show how the technical and pragmatic aspects of this modality (non-destructive, non-labelling, speed, relative low cost, etc.) enable it to be combined with more conventional methodologies. With this paper, we hope to open new avenues of research and extend the scope of available methodologies used in ecology and evolution.

Keywords: Raman spectroscopy; ecology; experimental evolution; phenotyping; pigment; vibrational imaging.

Figure 1.

A general overview of Raman spectroscopy. (a) A generalized overview of the optical set-up for a typical spontaneous Raman spectroscopic microscope. A monochromatic laser light illuminates the sample, emitting Raman scattering light. A small portion of this scattering passes through the objective lens and goes through the optical pathway to the dispersive polychromator (i.e. spectrophotometer), where it is captured by a CCD detector. (b) Raw spectral data must be processed through various procedures to improve the quality of the signal by removing noise (e.g. due to cosmic rays, the auto-fluorescence signal of the samples or from unavoidable technical variations). Because spectra are not always useful to visualize and are characterized by the molecular composition between each group or conditions, they are usually used as the input for multi-variate analyses such as principal component analyses, projection latent structure analyses, discriminant analyses and support vector machines, which are particularly powerful for identifying and discriminating individuals from the contribution of all or specific Raman wavelengths. (c) Representation of the inherent layers that contribute to the complex spectral signature.

Figure 2.

Perspective of microbial experimental evolution using Raman spectroscopy. Long-term cultures composed of single or multiple species are evolved in the presence or absence of varying levels of stressors. Population dynamics monitoring, as well as the characterization and isolation of emerging phenotypes (e.g. antibiotic-resistant mutants), could be performed by Raman spectral analysis, given the condition they exhibit different spectral markers (a specific pigment, amino acids, etc.), as suggested by recent studies [–9].

Figure 3.

Example of the advantages of Raman spectroscopy in evaluating the effects of food availability and parasitism on the secondary sexual characteristics in birds. In black birds, carotenoid pigments are an indicator of good health as their concentration is directly correlated with food availability and the immuno-resistance against parasitism [32]. Raman spectroscopy is known to provide specific signatures for pigment identification and quantification, as shown in the pioneering work of Thomas et al. [29,30], Galvan & Jorge [31] and Fernandes et al. [28], thus giving an alternative to the destructive and time-consuming conventional methods currently used by ecologists. This may become useful to monitor indirectly the health status of birds in natural populations. Likewise, the discrimination of intestinal and blood parasite species (copepoda, protozoa, etc.) could be done by spectral measurements in a systematic manner without the need for smears and hazardous manual identification. We show here a hypothetical example of linear discriminant analysis (LDA), in which three groups are identified (each point of the plot corresponds to a spectrum taken from one individual).