Viruses have been studied since 1892, and for many years the emphasis has been on the mechanism of viral infection and disease. By the 1950s, researchers had begun to use enzymes and genomes of viruses of bacteria (bacteriophages or phages) as tools in developing molecular biology. Therefore, phage biology was biased on a few models like Lambda, p22, and T4 among other families that are extremely abundant, diverse, and unexplored. Yet even in the well-studied phage models, new pieces of evidences suggest that structural proteins with unknown functions may be involved in trilateral interactions between phage, host, and environments. Recently, there has been renewed interest in phage therapy due to the development of antibiotic resistance and microbiome-related diseases require precision treatments. Nevertheless, critical features regarding the discovery of new phages, characterizing their unknown functions, and therapeutic applications remain unaddressed. This dissertation examines the issues listed above with various techniques on cultured and uncultured environmental phages. In chapter 2, unknown function open reading frames (ORFs) in marine virome were predicted and experimentally validated as a workflow to investigate the rapid growing Next Generation Sequencing (NGS) data. Machine-learning algorithms trained by protein features were utilized to detect and categorize phage structural genes. The selected candidates were cloned, over-expressed, and purified for reconstitutions of structures in vitro. Chapter 3 is dedicated to phage-glycan interactions in mucosal surfaces. Immunoglobulin-like (Ig-like) domains have been found in many phage structural proteins and they were proposed to bind glycans displayed on mucus subunits. In chapter 3, methods of phage engineering were tested for editing Ig-like domains. In addition, various assays were accessed for characterizing the phenotypes of phage interacting with mucus. Chapter 4 is a pioneer study of phage therapy targeting immature microbiota in the gastrointestinal systems. In collaboration with Center for Phage Technology at Texas A&M University, a toxin-free, non-replicating, non-lytic Bacillus prophage-like element, PBSX, was selected and engineered as “phagocin” aiming to kill pathogenic strains. In chapter 4, the DNA packaging pattern in PBSX was further investigated for better understanding of its life cycle and future applications.