How this works

Your sequence is first checked against the chosen rRNA's models (a reverse-complemented sequence is turned around, and long flanks such as a contig's are trimmed). It is then aligned with Infernal's cmalign to both the bacterial and the archaeal Rfam model, the same models every clade alignment on this site was built with. The alignment to the clade's own domain places each of your bases in that clade's alignment columns. Against Bacteria and Archaea pooled, the clade is counted in each domain's own columns and your bases come from the model your sequence fits better (higher bit score).

The clade is tallied one sequence per species, the top rRNA hit in its representative genome (or every rRNA copy in it, under Advanced settings). Type positions in E. coli, H. marismortui, Rfam consensus, or your own sequence's numbering.

Divergence is the Jensen–Shannon divergence between your base and the clade's A/U/G/C distribution (deletions and ambiguity codes are not counted, so a position where your sequence has a deletion has no divergence): near 0 where the clade agrees with you, 1 where no species in it shares your base. A conserved difference is a position where the clade is at least 90% one state (adjustable under Advanced settings) and your sequence has another. Bases your alignment places with a posterior probability below the cutoff in Advanced settings are left out, as they are for the clade.