PHAST is a web-based tool for rapidly and accurately identifying, annotating, and displaying prophage sequences in bacterial genomes or plasmids. It accepts raw DNA sequences or partially annotated GenBank files and performs database comparisons and phage feature identification to locate and annotate prophages. PHAST is up to 40 times faster and 15% more sensitive than other prophage identification tools. It can process raw DNA and GenBank files, provide detailed annotations, distinguish between intact and incomplete prophages, and generate downloadable, interactive graphics. PHAST is available at http://phast.wishartlab.com. Bacteriophage can be lytic or temperate. Temperate phages can integrate into the host genome, becoming prophages. Prophages are dormant and can be induced to excise themselves from the genome. Bacterial genomes can contain significant proportions of prophage genes, which may contribute to antibiotic resistance, environmental adaptation, and pathogenicity. Phages are also being investigated as potential antibiotics or cancer therapies. Prophage identification can be experimental or computational. Computational methods are now preferred due to the ease of sequencing bacterial genomes. PHAST uses a combination of sequence comparisons, tRNA and dinucleotide analysis, and hidden Markov scanning to identify prophage regions. It supports both raw and annotated genome data, and uses a custom phage and bacterial sequence database. PHAST can predict the completeness or potential viability of prophages and provides graphical outputs for both circular and linear genomic views. PHAST's performance was evaluated against a set of 54 manually annotated genomes. It achieved 85.4% sensitivity and 94.2% PPV using GenBank annotated files, and 79.4% sensitivity and 86.5% PPV using raw DNA sequences. PHAST is faster than other tools, with processing times of about 3 minutes for a typical genome. It is also more accurate than some existing tools. PHAST uses a combination of C and Java for its software, and a CGI framework for its web interface. It supports remote scripting and maintains a database of pre-computed bacterial genomes. PHAST's performance is superior to other tools, but some of its gains may be due to a larger phage sequence library and better keyword annotations. PHAST has some limitations, such as poor performance in identifying novel phages not closely related to its database. The DBSCAN algorithm used by PHAST assumes even distribution of phage-like hits, which may not always be true. PHAST may also split larger prophages into smaller ones due to a lack of BLAST hits. PHAST is a fast and accurate tool for prophage identification and annotation, producing results in minutes using raw or lightly annotated genome data. It provides extensive text summaries, downloadable figures, and interactive graphics. As phage and prophage databases continue to expand, PHAST's integrated approach is expected to improve its sensitivity and specificity.PHAST is a web-based tool for rapidly and accurately identifying, annotating, and displaying prophage sequences in bacterial genomes or plasmids. It accepts raw DNA sequences or partially annotated GenBank files and performs database comparisons and phage feature identification to locate and annotate prophages. PHAST is up to 40 times faster and 15% more sensitive than other prophage identification tools. It can process raw DNA and GenBank files, provide detailed annotations, distinguish between intact and incomplete prophages, and generate downloadable, interactive graphics. PHAST is available at http://phast.wishartlab.com. Bacteriophage can be lytic or temperate. Temperate phages can integrate into the host genome, becoming prophages. Prophages are dormant and can be induced to excise themselves from the genome. Bacterial genomes can contain significant proportions of prophage genes, which may contribute to antibiotic resistance, environmental adaptation, and pathogenicity. Phages are also being investigated as potential antibiotics or cancer therapies. Prophage identification can be experimental or computational. Computational methods are now preferred due to the ease of sequencing bacterial genomes. PHAST uses a combination of sequence comparisons, tRNA and dinucleotide analysis, and hidden Markov scanning to identify prophage regions. It supports both raw and annotated genome data, and uses a custom phage and bacterial sequence database. PHAST can predict the completeness or potential viability of prophages and provides graphical outputs for both circular and linear genomic views. PHAST's performance was evaluated against a set of 54 manually annotated genomes. It achieved 85.4% sensitivity and 94.2% PPV using GenBank annotated files, and 79.4% sensitivity and 86.5% PPV using raw DNA sequences. PHAST is faster than other tools, with processing times of about 3 minutes for a typical genome. It is also more accurate than some existing tools. PHAST uses a combination of C and Java for its software, and a CGI framework for its web interface. It supports remote scripting and maintains a database of pre-computed bacterial genomes. PHAST's performance is superior to other tools, but some of its gains may be due to a larger phage sequence library and better keyword annotations. PHAST has some limitations, such as poor performance in identifying novel phages not closely related to its database. The DBSCAN algorithm used by PHAST assumes even distribution of phage-like hits, which may not always be true. PHAST may also split larger prophages into smaller ones due to a lack of BLAST hits. PHAST is a fast and accurate tool for prophage identification and annotation, producing results in minutes using raw or lightly annotated genome data. It provides extensive text summaries, downloadable figures, and interactive graphics. As phage and prophage databases continue to expand, PHAST's integrated approach is expected to improve its sensitivity and specificity.