Team:UC San Diego/Model

Alaive

MODEL

We modeled the structure of CDR fragments using FREAD for 1,000,000 random amino acid sequences. FREAD is a software to predict loop structures de novo from existing databases [1]. We have constructed a database of 1,692 antibody sequence structures downloaded from the Protein Data Bank (PDB), which contained over 8,000 individual chains of amino acid sequences. Random sequences 9 amino acids long were used to replace the CDR3 sequence within a model structure of an antibody light chain: 1b6d. The loop structure of the new CDR3 sequence was predicted by comparing the new sequence to all similar sequences within the existing database. The predicted structure was encoded within a .pdb file and then docked to proteins found to be on T-cell surfaces using the tool Autodock Vina [2], which uses computation binding energy prediction to predict the binding orientation and strength between any two 3D molecular structures of peptides.

References

  • [1] Trott O, Olson AJ. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem 2010; 31: 455-461.
  • [2] Choi Y, Deane CM. FREAD revisited: Accurate loop structure prediction using a database search algorithm. Proteins 2010; 78: 1431-1440.