CMB2000/L19 Protein Structure Prediction Flashcards
What is the term given to multiple copies of the same protein?
Homomer
What is the term given to copies of different proteins?
Heteromers
What is the PDB?
Protein Data Bank
What is the accuracy of modern protein structure prediction algorithms?
> 80%
How does a protein structure prediction algorithm such as PSIPRED work? (2)
Uses information from alignments using iterative BLAST search
Extracts information using machine learning
Define machine learning.
An artificial neural network that has been trained on known secondary structure
Each amino acid in primary structure is given 1 of 3 which states?
Helix
Sheet
Coil
What are the 2 main approaches for predicting tertiary structure?
Template-based modelling
Ab initio modelling
Why is using a template to predict tertiary structure preferred? (2)
More straight-forward
Less computationally expensive
Give 1 downside of using a template to predict secondary structure.
Relies on similar protein having had its structure determined
How is the accuracy of protein structure prediction usually assessed?
By comparison to known structure
What is the CASP?
Critical Assessment of Structure Prediction
Done systematically since 1990s by international competition
Every 2 years since 1994
What is the most popular metric for assessing model quality?
Global Distance Test (GDT)
Describe the Global Distance Test (GDT).
All segments of 3, 5 and 7 amino acids from model superimposed to actual structure
Each are iteratively extended while they’re good enough
Good enough - distance between all residue pairs (represented by Ca atoms) is less than a threshold
What is the formula for GDT?
No. amino acids in segments/total number of amino acids in protein
