Complete data set.

Longitudinal monitoring of 50 honey bee (Apis mellifera) colonies from one Montana-based commercial beekeeping operation was conducted over the course of one year from November 2015 to October 2016. Samples were collected at six discrete time points, including two while colonies were located in California during (i.e. March 10, 2016) and just after almond pollination (i.e., April 23, 2016). By the end of the study in October 2016, colonies were sampled between three and six times. In total, we analyzed 262 samples, of which 37 were categorized as weak, 24 as average, 197 as strong, 4 were dead at the time of sampling. Honey bee colony population size was used as a proxy for colony health by counting the number of frames more than 2/3 covered with bees (i.e. weak (5 or fewer frames), average (6–8 frames), strong (9 or more frames) covered with bees)). Pathogen diagnostics was performed by PCR (1 = positive detection, 0 = not detected, NA—not assessed or no sample) for 13 pathogens and the most prevalent viral pathogens were assessed by qPCR. Relative RNA equivalents were determined by qPCR and were natural log transformed for statistical analyses (NA = not assessed by qPCR). Samples for which there was no virus detected by PCR would be below detection by qPCR and were thus considered as 0 during statistical analyses of qPCR. Over the course of the study, 22 colonies died, only four of which were collected as dead and analyzed (October 2016). The 18 other colonies that died throughout the study are included in the table at the time period they were found dead, then ommited from the table thereafter. (XLSX)