**Conflict of interest**

*Ficus citrifolia* and *Pisonia albida*, we can infer that they contribute with relevant functions in this soil system. The dwarfed tree species selected for this study influenced the structure and diversity of specific bacterial populations.

Amplicon-based pyrosequencing methods have major advantages over the tools that have been used in the past to study microbial community structure. Although the results presented in this chapter have a similar pattern as the results presented Rivera et al. (2018), it important to acknowledge certain biases that have been described for amplicon-based pyrosequencing. Even though 454 pyrosequencing has a higher resolving power than Sanger sequencing or EL-FAME analysis in 454 pyrosequencing, there are some sequencing errors and chimeras that can be retained in the datasets that can inflate the estimated richness of the sample. Bias can also occur with primer selection as the primers used can select for the most predominant DNA present in the sample underestimating the rare DNA in the sample [53]. Using inappropriate primers consequently can lead to questionable biological conclusions. Another concern is that the techniques used for processing amplicon pyrosequencing data can result in the detection of several hundred "false" OTUs, mostly at low abundance, rising the concern that species abundance can be overestimated [54]. More stringent techniques such as shotgun sequencing, Ion Torrent sequencing, and Illumina platforms have been developed that help mitigate some of the concerns with pyrosequencing, but these stringent technologies have

Soil bacterial communities have adapted to low rainfall at the Guánica Dry Forest; this could be a response to historical rainfall patterns encountered at the Guánica Dry Forest. The fact that 9 out of the 17 bacterial phyla identified were higher during the dry period supports this conclusion. For this forest, bacterial diversity did not change as a response to rainfall; however, equitability and richness changed demonstrating bacterial resilience. We are seeing how the same three bacterial phyla (*Proteobacteria*, *Actinobacteria*, and *Bacteroidetes*) are predominant during both dry and rainfall periods. Even though predominant bacterial phyla were the same during both periods under all tree species, differences were found at a finer scale. For instance, *Pisonia albida* had the soil with the most bacterial indicator species present. It is evident that this tree is shaping the soil microbiome in different ways. The general trend for predominant phyla found for Guánica is similar to the predominant phyla found in other terrestrial ecosystems even though the conditions of Guánica are unique. In the future, it would be nice to compare the sequences obtained in this study with other terrestrial environments but to the species level. This will help elucidate which are the species that could be playing

The project was funded by NSF Grant HRD-0734826 and is a contribution of the Centre of Applied Tropical Ecology and Conservation of the University of Puerto Rico. We appreciate the support of Mr. Larry Diaz, laboratory coordinator, and students from the Ecosystems Processes and Function laboratory of the University

important roles in ecosystem function and resilience.

**4.3 Potential disadvantages and bias with 454-pyrosequencing**

biases of their own.

*Microorganisms*

**5. Conclusions**

**Acknowledgements**

**56**

No conflict of interest.
