AUGUSTUS can be used as an ab initio program, which means that the prediction is based solely on the input DNA sequence. As a second possibility, the software may also incorporate hints on the gene structure coming from extrinsic sources such as BLAST [9] and DIALIGN [10] and [11] search results. With external hints, the prediction performance of AUGUSTUS can be increased significantly [12]. On the other hand, the computational costs to find external hints can exceed the costs of the program itself, in particular when large databases are used for BLAST search. Therefore, the search for external hints is not integrated into the AUGUSTUS web interface. In the following, we will focus on ab initio use of AUGUSTUS. In this case the command line program has two mandatory arguments: the query file and the species. The query file contains the input DNA sequences in uncompressed (multiple) FASTA [13] format. The species parameter denotes where the DNA sequences originate from, e.g. human, fruit fly, baker’s yeast, etc. There are several optional parameters which influence the way the genes are predicted. The output of AUGUSTUS is written to the command line output or into user-specified files. The output is compatible with the General Feature Format which can be visually interpreted by means of a genome browser [14]. Alternatively, the results can be transformed into gene maps in graphical representation via gff2ps [15]. A typical command line looks like: