Microorganisms and their chemical products are widely used as sources to isolate many drugs. To search for novel and
potential bioactive compounds from microorganisms, one approach is to acquire microbial samples from various
environments. However, with random collection and selection of the microbes, it would be hard to find the desired
bioactive compounds. To support the selection of the ecological habitat for collecting microorganisms in an efficient
way, we proposed a computational framework using molecular systematics and GIS-based modeling approaches. The
first step in this framework, molecular sequences and bioactivity profiles of microbes are used to build the phylogenetic
trees, whose leaf nodes are also associated with site location. Next, the phylogenetic diversity (PD) of
microbes/bioactivities among different geographic sites is estimated from the trees for the selection of interesting sites.
Using microbe occurrence and geographic data from the sites of interest, GARP algorithm is applied for the prediction of
species distribution in other areas. In addition, the PD values from each site are used in training data for prediction of
phylogenetic diversity and bioactivity diversity in unexplored areas.