Schematic representation of the ranking, according to the FROST score, of the template structures that are used to build the 6 models for a given query sequence (template structures are designated by the corresponding model name). S
represents the score above which there is less than 1% chance to find a false positive . Below this score we define the "twilight zone" of the method. In this zone the score does not allow us to predict with enough confidence whether the template structure is similar to the query sequence structure or not. DT1 is the first false positive, i.e., the template structure, not similar with the query structure, which has the highest FROST score. In this work the template structures we chose for ST2, DT1 and ST3 have comparable scores and they are found in the twilight zone. The template structure for ST1 has a score that is usually above S
. The ordering of the template structures for ST2, DT1 and ST3 can change from what is depicted on the figure depending on the query sequence.