Page 307 - Data Science Algorithms in a Week

288 Jose M. Prieto

Prediction of Antimicrobial Activities

Antibacterial and Antifungal Activity
Pioneering use of ANNs in microbiology has been quite restricted to the modeling
the factors contributing to microorganism growth (Hajmeer et al., 1997; Lou and Nakai,
2001; Najjar, Basheer, & Hajmeer, 1997) or yield of bioproducts (Desai et al., 2005).
QSAR studies of single chemical entities to shown the usefulness of artificial neural
network which seem to be of equal or somehow superior in prediction success to linear
discriminant analysis (García-Domenech and de Julián-Ortiz, 1998; Murcia-Soler et al.,
2004; Buciński et al., 2009).
Artificial intelligence also makes it possible to determine the minimal inhibitory
concentration (MIC) of synthetic drugs (Jaén-Oltra et al., 2000). Recently some works
have explored the use of such approach to predict the MIC of complex chemical mixtures
on some causal agents of foodborne disease and/or food spoilage (Sagdic, Ozturk & Kisi,
2012; Daynac, Cortes-Cabrera & Prieto, 2016).
Essential oils are natural products popularly branded as ‘antimicrobial agents’. They
act upon microorganisms through a not yet well defined mixture of both specific and
unspecific mechanisms. In this regard, ANNs are a very good option as they have been
successfully applied to processes with complex or poorly characterised mechanisms, as
they only take into account the causing agent and its final effect (Dohnala et al., 2005;
Najjar et al., 1997).
Indeed, the antibiotic activities of essential oils depend on a complex chemistry and a
poorly characterised mechanism of action. Different monoterpenes penetrate through cell
wall and cell membrane structures at different rates, ultimately disrupting the
permeability barrier of cell membrane structures and compromising the chemiosmotic
control (Cox et al., 2000). It is therefore conceivable that differences in the gram staining
would be related to the relative sensitivity of microorganism to Essential oils. However,
this generalisation on is controversial as illustrated by conflicting reports in literature.
Nakatani (1994) found that gram-positive bacteria were more sensitive to essential oils
than gram-negative bacteria, whereas Deans and Ritchie (1987) could not find any
differences related to the reaction. The permeability of the membrane is only one factor
and the same essential oil may act by different mechanisms upon different
microorganisms. As an example, the essential oil of Melaleuca alternifolia (tea tree)
which inhibited respiration and increased the permeability of bacterial cytoplasmic and
yeast plasma membranes, also caused potassium ion leakage in the case of E. coli and S.
aureus (Cox et al., 2001).
To further complicate matters, the evaluation antimicrobial activity of natural
products cannot be always attributed to one single compound in the mixture or when so,
the overall activity may be due to interactions between components of the essential oils.
In fact, synergism and antagonisms have been consistently reported as reviewed by Burt