RNAsalsa adapts known strcuture information to the considered sequences in several steps. Conserved structure features are verified via primary sequence variation and the detection of consistent and compensatory substitutions, which requires internal alignment steps. Remaining highly variable sections, where conserved structures cannot be detected, are folded by minimizing free energy. The result is a complete individual secondary structure for each sequence. A final alignment is done by taking both structure and sequence information of each position into account. For further tree reconstruction analysis, a sequence alignment with a consensus structure allows the application of RNA substitution models in the conserved paired regions of the considered RNA genes. It is important to notice, that RNAsalsa uses structure information for adjusting and refining the sequence alignment and vice versa.