Bacterial influences on harmful algal blooming Alexandrium populations in the Nauset Marsh System natural laboratory: The occurrence and severity of harmful algal blooms produced by dinoflagellates such as members of the genus Alexandrium that are responsible for Paralytic Shellfish Poisoning outbreaks is increasing worldwide. Harmful algal blooms not only present a threat to human health via consumption of contaminated shellfish but can also have devastating impacts on local economies dependent on the shellfish industry. We have been studying the Nauset Marsh System on Cape Cod, Massachusetts, USA that experiences annual blooms of Alexandrium fundyense associated with Paralytic Shellfish Poisoning outbreaks that have increased in intensity in the system over the last 20 years. Two kettle holes or salt ponds within the Nauset Marsh System experience consistent localized blooms typically offset in timing and thus present a "natural laboratory" for studying the diversity and dynamics of A. fundyense blooms. Many ecological drivers of these blooms remain unknown and in particular, little is known about the relationship between the Nauset Marsh System bloom dynamics and the microbial communities associated with A. fundyense. In this study we tested to what extent A. fundyense blooms: (i) co-occur with specific types of bacterial populations across samples and ponds, and, (ii) contain identifiable indicator OTUs that precede the onset of blooms and may thus serve as predictors. Our project hypothesis is that abrupt, short-term changes in A. fundyense population dynamics and structure are associated with identifiable changes in abundance and community composition of bacteria. To test this hypothesis, we collected microbial community structure data from the 2010, 2011 and 2012 blooms. Our highest-resolution data (every other day at both high tides) were collected before, during, and after A. fundyense blooms in Mill and Salt Ponds of the Nauset Marsh System during the spring of 2012 using a combination of manual sampling and collection via an automated water sampler prototype equipped to preserve samples for downstream DNA analysis. We employed 454 or Illumina next-generation DNA amplicon sequencing of either the V6-V4 or the V6-targeted bacterial SSU rRNA gene hypervariable regions on 332 of the resulting samples to generate bacterial community profiles. Our preliminary data present a compelling case for linkages between bacterial community dynamics and those of A. fundyense populations but modeling-based approaches are required to add predictive value to our approach. Overall community similarity was moderately high but still showed discreet communities, particularly for the pre-bloom and post-bloom periods. Furthermore, the peak- and post-bloom populations in Mill Pond were similar to those in Salt Pond despite the one-month difference in bloom onset. Application of newly designed algorithms and improved methods to detect co-occurrence patterns among bacterial OTUs will determine whether true bacterial indicators of bloom dynamics exist.