On the reliability and the limits of inference of amino acid sequence alignments

Details of domain pairs used in the manuscript

• Set of Million domain pairs

• List of million domain pairs
• Subset of million domain pairs whose Δ_optimal > 0
• Subset of million domain pairs whose Δ_optimal ≤ 0 < Δ_marginal
• Subset of million domain pairs whose Δ_marginal ≥ 0

• Five distinct groups of domain pairs based on SCOP hierarchy

• List of domain pairs sampled at the same family level
• List of domain pairs sampled at the same superfamily level
• List of domain pairs sampled at the same fold level
• List of domain pairs sampled at the same class level
• List of domain pairs sampled at different class level

Expected and Empirical inter-alignment distance vs Markov time plots

Note: All the reference alignments reported by the 3 structure alignment programs are canonicalized in the form D*I* before computing the distance measures. See supplementary materials for more information.

• Set of Million domain pairs

• Expected inter-alignment distance vs Markov time plots for Million domain-pairs
  Substitution models: BLOSUM MMLSUM VTML PAM | Structure alignment programs: TM-Align MMLigner Dali
• Empirical inter-alignment distance vs Markov time plots for Million domain-pairs
  Substitution models: BLOSUM MMLSUM VTML PAM | Structure alignment programs: TM-Align MMLigner Dali
• Expected and Empirical inter-alignment distance vs Markov time plots after breaking down by their compression statistics for BLOSUM model
• Expected and Empirical inter-alignment distance vs Markov time plots after breaking down by their compression statistics for MMLSUM model
• Expected and Empirical inter-alignment distance vs Markov time plots after breaking down by their compression statistics for VTML model
• Expected and Empirical inter-alignment distance vs Markov time plots after breaking down by their compression statistics for PAM model


• E-value vs Markov time plot for Million domain-pairs
  Substitution model: BLOSUM | Structure alignment program: MMLigner

• Five distinct groups of domain pairs based on SCOP hierarchy

• Expected inter-alignment distance vs Markov time plots
  Substitution models: BLOSUM MMLSUM | Structure alignment programs: MMLigner