BLAST stands for Basic Local Alignment Search Tool.
The purpose of BLAST is to help researchers find and compare similar gene sequences among nucleotides or proteins and narrow them down by statistical significiance from a library of gene sequences. What this does is it helps identify similar members of gene families that have evolved throughout time. It was designed by members of the NIH, Stephen Altschul, Warren Gish, Webb Miller, Eugene Myers, and David J. Lipman and published in 1990 in the Journal of Molecular Biology.  The process of how BLAST is used is that it uses short matches between sequences and filters out based on initial words, a process called seeding. It then looks for commonality between the sequences and matches them up. There are 7 different programs that BLAST comes from, each program specializes in a certain type of BLAST you want to run. There are nucleotide to nucleotide BLAST, a protein to protein BLAST, position to specific iterative BLAST, Nucleotide 6-frame translation-protein, nucleotide 6-frame translation-nucleotide 6-frame translation, protein-nucleotide 6-frame translation, or large numbers of query sequences (megablast). [1,2]
Many researchers in bioinformatics, genetics, genomics, and metagenomics use this tool to their advantage. It's an easy and fast way to obtain data to compare them among species. For example, a researcher studying a protein gene in Plasmodium falciparum may want to BLAST the protein sequence to see if humans have a similar gene which could potentially be studied to invent a drug target for malaria. Usually BLAST can be the first stop researchers make to expand more information about their subject's species. Other things it can be used for is in DNA mapping, identifying a phylogeny, and locating domains. [1,2]
How to use:
To use BLAST, the input format are sequences that are in FASTA or Genbank format and weight matrix. The output method can be chosen, either HTML, plain text, and XML.