Aplicación de algunas técnicas de estadística multivariada al estudio de la conservación in vitro de germoplasma de Dioscorea alata L.

Rational use of multivariate statistics, modeling and biological knowledge can help the researcher to design a tested crucial experiment and to reach consistent conclusions. This work based on this statement and on a previous report of in vitro conservation of D. alata . It aims to apply some multivariate statistical techniques to determine the main factors which contribute to differentiate and to group the treatments. The most suitable factors for the preservation of genetic resources of this species were also defined. The D. alata conservation process was evaluated by principal component analysis at 9 months. Percentage of survival, phenolization, explants with leaves, rooting, sprouting and leaf senescence, stem length, root number and number of novo nodes per explants were the analyzed variables. Results showed that the survival rate variables, percentage of explants with leaves, stem length, number of novo nodes and number of roots express much of the observed variation in the processes of growth and development. On the other hand, deterioration and aging of in vitro nodal segments buds is better explained by the percentage of phenolization variables and leaf senescence. The best treatments for in vitro conservation of D. alata based on variables with higher contribution in the multivariate statistical analysis, corresponded to the culture medium D - 571 with the 1.5% mannitol, 0.1 or 1 mg.l -1 BA and 2 g.l -1 of activated charcoal. Keywords : activated charcoal, benzylaminopurine, cluster analysis, principal components, mannitol, yam