Toxins produced by cyanobacterial blooms in freshwater lakes are a serious public health problem. The conditions leading to toxin production are unpredictable, thereby requiring expensive sampling and monitoring programs globally. We explored the potential of volatile organic compounds (VOCs) to indicate microcystin presence and concentration and microbial community composition in Upper Klamath Lake, Oregon. Elastic net regularization regression selected 24 of 229 detected m/z + 1 values (corresponding to unique VOCs) in models predicting microcystin contamination. These models outperformed regression models based only on environmental parameters, including chlorophyll, pH, and temperature. Several m/z + 1 values selected by elastic net were putatively identified as saturated fatty aldehydes, which defend cyanobacteria against oxidative stress. Elastic net also identified unique sets of m/z + 1 values that predicted the relative abundance of the dominant bacterial phyla, classes, and cyanobacterial genera. VOCs appear to reveal the physiological status of cyanobacteria during toxic blooms and may be a key component of lake monitoring strategies. IMPORTANCE Harmful algal blooms are among the most significant threats to drinking water safety. Blooms dominated by cyanobacteria can produce potentially harmful toxins and, despite intensive research, toxin production remains unpredictable. We measured gaseous molecules in Upper Klamath Lake, Oregon, over 2 years and used them to predict the presence and concentration of the cyanotoxin, microcystin, and microbial community composition. Subsets of gaseous compounds were identified that are associated with microcystin production during oxidative stress, pointing to ecosystem-level interactions leading to microcystin contamination. Our approach shows potential for gaseous molecules to be harnessed in monitoring critical waterways.