Day 2: Prediction

DAY 2 (10.11.16):
Protein structure prediction
  • Protein structure prediction is the inference of the secondary and tertiary (3D) structure of a protein given the input primary sequence. 
  • To speed up protein folding in a computer, one can "cheat" by taking fragments or chunks from existing proteins in the PDB (Protein Data Bank) and recombining them to generate protein structures. 
    • When short fragments (3 to 9 amino acids long) are used, this is referred to as "Abinitio" or "denovo" structure prediction.
    • When large fragments (or "templates") are used, this is referred to as "Homology" or "Comparative" modeling.
  • Instead of taking random fragments from the PDB, these are instead biased by sequence similarity, predicted secondary structure and solvent accessibility. If statistically significant LARGE fragments can be detected, CM protocols are run, otherwise AB is used.
ć
Sergey Ovchinnikov,
Nov 2, 2016, 9:55 AM
ć
Sergey Ovchinnikov,
Nov 2, 2016, 10:11 AM
ć
Sergey Ovchinnikov,
Nov 2, 2016, 10:01 AM
ċ
t574_hybrid.pse
(955k)
Sergey Ovchinnikov,
Nov 2, 2016, 10:02 AM