Genetic screens are powerful tools for functional genomics. I developed a method to substantially increase the throughput of genetic screens in Escherichia coli and applied it to study bacteria–bacteria and bacteria–drug interactions. To investigate metabolic cross-feeding in microbial communities, I systematically screened single-gene knockouts and found that strains benefiting from shared goods were typically defective in amino acid, nucleotide, or vitamin biosynthesis, or in central carbon metabolism. Pairwise experiments identified vitamin auxotrophs as optimal cross-feeding partners. In larger assemblies of auxotrophs, consortia rapidly coalesced around vitamin-deficient strains, stabilized by multiple cross-feeding interactions. These results suggest vitamins are ideal shared goods by supporting community growth while maintaining member interdependence. To explore the antibacterial activity of non-antibiotic drugs, I screened 200 compounds using a barcoded E. coli knockout library. Network analysis of drug–drug similarities revealed that antibiotics clustered into modules corresponding to known mechanisms of action, whereas non-antibiotics largely remained unconnected. Nonetheless, about half of non-antibiotics formed distinct clusters, indicating shared and potentially unexploited antibacterial targets. Additionally, analysis of efflux systems revealed they impact antibiotics and non-antibiotics alike, raising concerns that non-antibiotic exposure may promote cross-resistance in vivo. Based on these findings I designed a custom library and machine learning pipeline for mechanism-of-action discovery. My studies demonstrate the versatility of high-throughput genetic screening for investigating microbial ecology and pharmacology. By advancing both methodological and biological understanding, this work highlights how genetic screens can uncover principles of cooperation in microbial communities and illuminate the under-characterized antibacterial effects of non-antibiotic drugs.