In my group we work on high-throughput discovery of toxin-antitoxin systems in microbial genomes. Our work in this area is high impact, being published in PNAS and MolCell, and we are supported by a generous KAW project grant. The breakthrough and availability of AlphaFold2 that has happened this year is excellent time for us, as as many of the the toxins and antitoxins have novel structural folds. Our plan is not only to predict the structures of novel toxins and antitoxins, but to use AlphAFold2 to predict the interaction surfaces of toxin-antitoxins, which we can validate with mutagenesis. Associated with this is comparative analyses of thousands of genomes and proteomes. To do this on the scale that we wish to, we need access to additional computational power.