Top 16 pfam domains with least selective active drugs.

Only domains present in the H. sapiens proteome are shown, and are sorted and selected by decreasing domain selectivity among FDA approved drugs. We also excluded domains with under 100 amino acid residues in the homology model, to avoid small repeats and domains unlikely to be drug targets. Additionally, we report for multi-domain clusters only the most frequent and functionally descriptive members as outlined in the Methods section (see “De-duplication of Single Domain Clusters”), and also removed domains of unknown function (DUFs). To the left of the slash in each column is the median target selectivity of FDA approved compounds active against this domain, and to the right is the median selectivity of non-FDA compounds. Selectivity is quantified by three methods of increasingly grouped protein targets as described in the text. For example, the non-FDA approved compounds active against targets with the PF00183 domain are active against a median of 12 targets with distinct GenBank identifiers, so this domain has a 12 after the slash in the first column.