Sequence variation in ligand binding sites in proteins

1
Sequence variation in ligand binding sites in
proteins BMC Bioinformatics 2005, 6240 Thomas J
Magliery, Lynne Regan
2
Background

• Residues that are conserved (not conserved) in
  MSA are considered to be important (not
  important)
• It is not applicable to families of structurally
  similar proteins that bind diverse ligands
• This works demonstrates that sometimes binding
  sites can be predicted by detecting
  hypervariation in MSA

3
Statistical Free Energy

• Introduced by Lockless and Ranganathan
• (to study coupled mutations)

Pi probability to have residue i on a particular
position in MSA compared to reference MSA (Pref)
Root mean square average log for all 20 residues
SFE
4
Multinomial probability

• N total number of sequences
• nx number of sequences with amino acid x at the
  given position
• fx expected frequency

For all 34 positions in the TPR motive (both
dependent on N)
5
Entropies

• Sequence (Shannon) entropy
• px proportion of sequences with amino acid x at
  position i

Relative entropy Distance between two
distributions
6
TPR-binding site
Tetratrico peptide repeat
7
TPR-binding site
8
Sample size
9
The ankyrin repeat
10
Zinc fingers
11
PDZ domain
X-(S/T)-X-(V/I/L)-CO2 X-Phob-X-Phob-CO2
12
Conclusions
The method can be useful for analysis
functionally diverse protein families.
13
Conclusions

• The method can be useful for detecting
  functionally important residues
• Can be applied only for a set (gt100) of proteins
  that have the same scaffold but diverse functions
• Can be a supplement to the existing methods that
  analyze residue conservation