Pseudomonas aeruginosa (Pa), a common pathogen, causes a wide range of diverse infections that lead to enormous health care costs and unacceptably high morbidity and mortality. Pa drug resistance is an increasing problem. Thus, new Pa infection prevention, elimination and mitigation strategies are needed. In many infection contexts, including biofilms, infection sites, and multispecies communities, Pa is often in microoxic environments and microoxia imposes unique challenges to Pa energy generation, catabolism, and biosynthesis, signaling, and stress responses among other processes. Yet, microoxic fitness determinants are poorly understood. To enhance our ability to identify genes involved in microoxic fitness and conditions that will best reveal microoxic fitness phenotypes, we will leverage our prior work on the analysis of transcriptional signatures using publicly available data. Pa has been extensively studied using transcriptomics with over 4000 whole transcriptome microarray and RNA-seq datasets in public repositories. The Hogan Lab in collaboration with the laboratory of Dr. Casey Greene (University of Colorado, Anschutz) has used these transcriptome datasets to create and validate normalized compendia for cross-experiment comparisons and developed machine learning approaches to facilitate the use of these data to identify genes with correlated expression patterns. Using these resources, we can observe robust co-expression patterns that may not stand out in single experiments. We found that known microoxic fitness genes are significantly correlated in expression patterns with each other and with a large set of uncharacterized genes. In Aim 1, we will construct mutants in candidate microoxic fitness genes, analyze their fitness in microoxic and normoxic conditions, and compare the phenotype observed using single mutants to the results from a Tn-Seq performed in similar conditions. In light of our data that show that medium composition and strain background can influence the expression level of microoxic fitness genes even when cell density and O2 availability are constant, in Aim 2, we will use transcriptome compendia to identify conditions in which microoxic fitness genes are most highly expressed. We will use those conditions to assess the phenotypes of mutants in candidate microoxic fitness genes. Prioritized hits will be validated by complementation and further characterized in future work. Upon completion of these studies, we will have expanded our knowledge of microoxic fitness determinants, and we will have tested a strategy for leveraging public transcriptomics data for identifying candidate genes and assay conditions that might be broadly useful in other systems and for other processes.