 This study investigates message-level drivers of the popularity and virulity of tweets about COVID-19 vaccines using machine-based text-mining techniques, examining topic communities of the most liked and retweeted tweets using network analysis and visualization. The results show that topics related to vaccine development and people's views, and vaccine efficacy and rollout, positively predicted likes and retweets. Network analysis revealed that these topics clustered around the 2,500 most liked and most retweeted tweets. The study suggests that addressing these topics in vaccine campaigns, could help the diffusion of content on Twitter. This article was authored by Jimin Jong, E-Wong, Molasher and others.