认识生态未知物:用统计系统方法研究生物多样性

包括变性梯度凝胶电泳实验、16S rDNA克隆库、DNA生物芯片在内的分子生物学技术表明在自然环境中物种存在极端多样性。这些种类绝大多数都是不可培养的，不能完全了解其特征。通常运用Shannon和Simpson的经验参数来评估多样性。这些参数依赖于样品的大小r不可靠。目前3种有效途径，即参数评估、非参数评估和系统多样性评估来解决生物多样性问题，同时和分子学技术相结合，三者可以增加种群数以及多样性评估的正确性，并可以在不同群落间进行严格比照。

Knowing the Unknown: Statistical Approaches towards Understanding Microbial Diversity

Molecular techniques, such as denaturing gradient gel electrophoresis (DGGE), 16S rDNA clone library, and DNA microarray, reveal extraordinarily diverse species in the natural environment. Most of these species are unculturable and cannot be characterized precisely. Empirical indices like Shanon and Simpson have been traditionally used to evaluate the diversity. These indices, however, have certain disadvantages because of their dependence on sample sizes. To overcome such disadvantages, three approaches have been developed, which include parametric estimation, nonparametric estimation, and phylogenetic diversity estimation. These methods when integrated with molecular techniques, can enhance accurate estimation of species richness and diversity, and allow rigorous comparisons of different communities.