Inferring repeat-protein energetics from evolutionary information
Fig 4
Most favorable simulated sequences have very similar repeats, yet they are different to the consensus repeat.
On A, we plot the energy vs. the identity between the repeats that constitute the sequence. Even though the deviation is large, most stable sequences tend to have more similar repeats. On B, we plot the energies of simulated sequences vs the identity to the consensus of the family. In all cases, the identity to the consensus is small and uncorrelated to the energy, indicating that sequences which differ significantly from the consensus can be stable variants of the family.