Multivariate analysis in phytopathology: Options and opportunities in data mining to face new molecular information

BALZARINI, M.; BRUNO, C.; FERNÁNDEZ, E.

Phytopathology in the Omics Era

Research Signpost

Lugar: Kerala; Año: 2011; p. 1 - 20

Multivariate techniques and machine learning methods are used by phytopathology scientists working with highly dimensional genomic and proteomic data. Molecular data are often obtained by high throughput technologies, so the amount of data generated is huge and it demands an overall process of finding and interpreting patterns from data where statistical and bioinformatic techniques are crucial. When properly analyzed, molecular marker data allow biological meaningful reflection of DNA and RNA comparisons between individuals and populations. An accurate interpretation of such type of omic data depends on understanding the strengths and limitations of the available analytic methodologies for processing molecular information. In this chapter we describe a number of statistical methods, multivariate in nature, used with molecular data for ordination and classification of molecular information in reduced spaces, give brief illustration of applications in Phytopathology, and discuss how they can be used in complementary manner no only in cross-sectional but also in spatial genomics. We talk about methods of ordination and classification that were too difficult to implement in the near past because their high computing demand but which are now easily to use due to the increase in computing capabilities.