Armed with these methods the algorithm will start off treating the chain as a single segment domain and divide it using Method 1 or 2, whichever yields the higher split value. If a two segment domain is found at any point, it is scanned using Method 3. The algorithm continues to divide the protein, until it is checked by one of the constraints. Constraints are described in the next section and are also present to allow the algorithm to be flexible and fast.
Note that none of these methods will deal with domains consisting of three or more segments. Domains such as these are not dealt with explicitly in the algorithm at the scanning stage as the complexity of the scan would increase rapidly. However, they are allowed implicitly at a later stage (described below). Such domains are found to be quite rare, making up only a small fraction of the total number of domains in the database.