Hybrid use of Raman spectroscopy and artificial neural networks to discriminate Mycobacterium bovis BCG and other Mycobacteriales.

Even in the face of the COVID-19 pandemic, Tuberculosis (TB) continues to be a major public health problem and the 2nd biggest infectious cause of death worldwide. There is, therefore, an urgent need to develop effective TB diagnostic methods, which are cheap, portable, sensitive and specific. Raman spectroscopy is a potential spectroscopic technique for this purpose, however, so far, research efforts have focused primarily on the characterisation of Mycobacterium tuberculosis and other Mycobacteria, neglecting bacteria within the microbiome and thus, failing to consider the bigger picture. It is paramount to characterise relevant Mycobacteriales and develop suitable analytical tools to discriminate them from each other. Herein, through the combined use of Raman spectroscopy and the self-optimising Kohonen index network and further multivariate tools, we have successfully undertaken the spectral analysis of Mycobacterium bovis BCG, Corynebacterium glutamicum and Rhodoccocus erythropolis. This has led to development of a useful tool set, which can readily discern spectral differences between these three closely related bacteria as well as generate a unique spectral barcode for each species. Further optimisation and refinement of the developed method will enable its application to other bacteria inhabiting the microbiome and ultimately lead to advanced diagnostic technologies, which can save many lives.