Culture-independent analyses of host-microbial communities can be used in lieu of Koch's postulates for an uncultured pathogen as we have done for citrus greening disease. Prior to our work on this subject, there was strong evidence that the causal agent of citrus greening disease was a phloem-limited, uncultured alphaproteobacterium, <i>Ca</i>. Liberibacter asiaticus (CLas). The vast majority of plants with symptoms were PCR-positive for CLas. However, as this organism has not been cultured, Koch's postulates have never been performed. There were some suggestions that disease may be caused by other bacteria or bacteria in combination with CLas. Subsequent metagenomic analysis of infected phloem tissue showed that the only bacterium present in citrus phloem was CLas. Where a causal agent is not known, careful experimental design can identify an organism strongly associated with disease. In addition, the associated organism may not be the exclusive causal agent of disease. That is, a set of bacterial functions commonly shared by several taxa may cause disease. In human, animal, or plant microbiome research, it is important to keep in mind the many environmental factors that confound the results. We propose that the number of environmental confounders associated with disease be minimized in experimental design. Examples will be presented where culture independent analyses identified organisms correlated with the presence of disease and land use.