Multivariate statistics for analysis of honey bee propolis

Alghamdi, Abdulaziz Saleh H and Gray, Alison (2017) Multivariate statistics for analysis of honey bee propolis. In: Royal Statistical Society Conference 2017, 2017-09-04 - 2017-09-07.

[img]
Preview
Text (Alghamdi-Gray-RSS-2017-Multivariate-statistics-for-analysis-of-honey-bee)
Alghamdi_Gray_RSS_2017_Multivariate_statistics_for_analysis_of_honey_bee.pdf
Final Published Version

Download (582kB)| Preview

    Abstract

    Honey bees play a significant role ecologically and economically, through pollination of crops. Additionally, honey can be considered as one of the finest products of nature, with a wide range of beneficial uses, including use in cosmetic treatment, eye diseases, bronchial asthma and hiccups. Honey bees also produce beeswax, royal jelly and propolis. Propolis is a resinous bee product, which consists of a combination of beeswax and resins which have been gathered by honey bees from the exudates of various surrounding plants. It is used by the bees to seal and maintain the hives, but is also an anti-infective substance which may protect them against disease. Propolis possesses a highly resinous, sticky gum appearance and its consistency changes depending on the temperature. It becomes elastic and sticky when warm, but hard and brittle when cold. Furthermore, its colours vary from yellowish-green to dark brown, depending on its age and source. The purpose of this research is to use statistical analysis to study the biochemical properties of propolis, which have attracted much attention. Biochemical analysis of propolis leads to highly multivariate metabolomics data. The main benefit of metabolomics is to generate a spectrum, in which peaks correspond to different chemical components, making possible the detection of multiple substances simultaneously. Relevant spectral features may be used for pattern recognition. This work will investigate the use of different statistical methods for analysis of metabolomics data from analysis of propolis samples using Mass Spectrometry (MS). Methods studied will include pre-processing methods and multivariate analysis techniques such as principal component analysis (PCA), clustering methods and partial least squares (PLS) methods. Background material and initial results of data analysis will be presented from samples of propolis from beehives in Scotland.