**4. PCA applied to the interpretation of quantum chemical calculations**

Quantum chemical calculations in large flexible molecules represent a challenge, due to the high number of combinations of conformationally relevant parameters and computational resources/capabilities required. This is an extremely complicated task, unless a systematic approach is carried out. In this section, a PCA-based methodology allowing the correlation between molecular structures and properties of a conformationally flexible molecule (arbutin) is described. This procedure is simple and requires relatively modest computational facilities (Araujo-Andrade et al. 2010).

Arbutin is an abundant solute in the leaves of many freezing- or desiccation-tolerant plants. It has been used pharmaceutically in humans for centuries, either as plant extracts or, in more recent decades, in the purified form. Arbutin acts as an antiseptic or antibacterial agent on the urinary mucous membranes while converting into hydroquinone in the kidney 36 Principal Component Analysis

The same approach performed for bacteria/Cd+2 interaction was employed to analyse the interaction of the other metal ions with the bacterial surface. The scores and loadings plots obtained after the removal of outliers from PCA, carried out on the spectral data sets corresponding to samples of different concentrations of the three metal ions attached to the bacterial surfaces are shown in Fig. 8. A clear discrimination between objects corresponding to different concentrations of lead and zinc trapped on the bacterial surface was observed along the PC-space [Figure 8(a) & (b)]. The information provided by the loadings explains the distribution of samples in clusters surrounded by ellipses and labeled as I, II and III.

The dashed rectangles depicted in the loadings plots indicate the spectral regions that are statistically influent in the distribution of objects. For lead, region labeled as I corresponds to the superficial polysaccharides (1000-800 cm-1 region), region labeled as II, to the fingerprint region (700-250 cm-1), and region labeled as III, to the amide I region (1650-1500 cm-1). For zinc, region I corresponds to superficial polysaccharides (1000-750 cm-1), region II, to amide I

It is important to point out that in both bacteria/Pb+2 and bacteria/Zn+2 interactions, the same spectral regions were identified as influent for the objects distribution observed in PCA. This indicates that similar molecular structures may be involved in the bacteria/metal interaction. The scores plot corresponding to the bacteria/Ni+2 samples indicate a poor discrimination among objects [Fig. 8 (c)]. According to the loadings plots of the bacteria/Ni+2 samples, the spectral similarities and differences between samples are mainly

From the results discussed and described above, it can be concluded that Raman spectra allow obtaining chemical information related with bacteria/metal interactions, and also with metal ion concentrations. PCA allowed an efficient elucidation of the information obtained from the spectra. Furthermore, the one-dimensional loadings analysis allowed identifying the influent spectral regions, as well as the molecular structures involved in the

Quantum chemical calculations in large flexible molecules represent a challenge, due to the high number of combinations of conformationally relevant parameters and computational resources/capabilities required. This is an extremely complicated task, unless a systematic approach is carried out. In this section, a PCA-based methodology allowing the correlation between molecular structures and properties of a conformationally flexible molecule (arbutin) is described. This procedure is simple and requires relatively modest

Arbutin is an abundant solute in the leaves of many freezing- or desiccation-tolerant plants. It has been used pharmaceutically in humans for centuries, either as plant extracts or, in more recent decades, in the purified form. Arbutin acts as an antiseptic or antibacterial agent on the urinary mucous membranes while converting into hydroquinone in the kidney

**4. PCA applied to the interpretation of quantum chemical calculations** 

(1700-1500 cm-1), and region III, to the fingerprint region (450-250 cm-1).

in the fingerprint region (region surrounded by a blue dashed line).

objects/samples distribution in the scores plots.

computational facilities (Araujo-Andrade et al. 2010).

**3.3.3 PCA results for other metal ions** 

**3.4 Final remarks** 

(Witting et al., 2001). It is also used as a depigmenting agent (skin whitening agent) as it inhibits melanin synthesis by inhibition of tyrosinase activity.

From a chemical point of view, arbutin is a flexible molecule composed by a glucopyranoside moiety bound to a phenol ring (Fig. 9). It has eight conformationally relevant dihedral angles, five of them related with the orientation of the hydroxyl groups and the remaining three taking part in the skeletal of the molecule.

Up to our knowledge, no attempts to use of a PCA based methodology for the structural analysis of quantum chemical information were reported.

Fig. 9. Arbutin molecule, with atom numbering scheme. (copyrighted from Araujo-Andrade et al., 2010)
