Figure 1

Logic Flow diagram of the ReformAlign alignment strategy. The algorithm creates an initial profile based on a starting alignment deriving from an established aligner. Then, all the sequences are aligned against the formed profile to obtain the reformed alignment. Notice that in this step, in case a new insertion is detected, the algorithm automatically switches to a profile fine-tuning mode in order to modify the profile to also account for the new insertion(s). After all sequences have been successfully aligned against the profile, the deriving alignment is indirectly inferred by merging all the estimated subalignments into a unified set and it is then freed from uninformative void columns (i.e. columns consisting exclusively of gaps). The process is repeated using the reformed alignment as starting alignment until the alignment between two successive runs remains unchanged or a predefined maximum number of iterations is reached.