Transcript PROTEIN STRUCTURE CLASSIFICATION

PROTEIN STRUCTURE
CLASSIFICATION
SUMI SINGH
(sxs5729)
Levels of Protein Structure
2
Traditional Architecture
WOOD, BRICK etc.
Form
fits
function
material/building blocks
Molecular Architecture
AMINO ACIDS
*****************************************************************
Number of Amino Acids found in Eukaryotic Proteins=
20 (found in universal genetic code)+ 2 (synthetically incorporated therefore not included in discussion)
Possible number of protein sequence of size 300 =
20300
This number is greater than the total number of atoms in the
universe
Thanks to: Frank Lloyd Wright for graphics
3
Evolution
Evolution has selected a very small subset of those
protein sequences
< 30,000 in humans and an even smaller number of
protein structures (1000–5000)
Ratio– 1:6
Conserved structures are expected to reflect
functional similarities (interaction with other
molecules)
4
Why Compare Protein Structures?
Sequence
Structure
Function
 Low sequence similarity may yield very similar structures
 Sometimes high sequence similarity yields different structures
5
Know your dataset
FOR THE CURRENT PROJECT
PDB: Protein Data Bank
•
The Protein Data Bank (PDB) archive is the single worldwide repository
of information about the 3D structures of large biological molecules,
including proteins and nucleic acids.
•
These are the molecules of life that are found in all organisms including
bacteria, yeast, plants, flies, other animals, and humans. Understanding
the shape of a molecule helps to understand how it works.
•
This knowledge can be used to help deduce a structure's role in human
health and disease, and in drug development. The structures in the
archive range from tiny proteins and bits of DNA to complex molecular
machines like the ribosome.
•
Web address: http://www.rcsb.org/pdb/home/home.do
SCOP Dataset
“Nearly all proteins have structural similarities with
other proteins and, in some of these cases, share a
common evolutionary origin. The SCOP database,
created by manual inspection and abetted by a
battery of automated methods, aims to provide a
detailed and comprehensive description of the
structural and evolutionary relationships between
all proteins whose structure is known. As such, it
provides a broad survey of all known protein folds,
detailed information about the close relatives of
any particular protein, and a framework for future
research and classification.”
• http://scop.berkeley.edu/
Starting at the bottom, the hierarchy
of SCOP domains comprises the
following levels
-- Species representing a distinct
protein sequence.
-- Protein grouping together similar
sequences of essentially the same
functions.
-- Family containing proteins with
similar sequences but typically
distinct functions.
--
Superfamily bridging together
protein families with common
functional and structural
features inferred to be from a
common evolutionary ancestor.
-- Levels above Superfamily are
classified based on structual
features and similarity, and do not
imply homology:Folds grouping
structurally similar superfamilies.
Structural
Fingerprints/Features
Structure comparison is an NP-Hard problem.
There are no fast structural alignment algorithms
that can guarantee optimality within any given
similarity measure. Therefore, existing structure
comparison methods employ heuristics.
There are different approaches for extracting
structural features.
We use Triangular Spatial Relationship to generate
keys.
HUH!! LOOKS LIKE I HAVE DONE
EVERYTHING..
SO WHY ARE WHY AM I HERE?
FOR THE CURRENT PROJECT
What do you get from me?
• A file of Keys Created Representing each
Protein Structure.
• Each of these files of keys representing
protein has been correctly classified into their
respective Superfamilies.
• That will give you the class information for the
files. It is a hypothesis that each file belonging
to same class must have similar keys. You must
be able to test this hypothesis.
Biggest Challenge
COMBINATORIAL EXPLOSION
30,000,000.00
25,000,000.00
Key Count
20,000,000.00
15,000,000.00
10,000,000.00
5,000,000.00
4,410,549.00
0.00
0
50
100
150
200
250
300
350
Number of Amino Acids per Protien
400
450
500
550
600
For the current project
• Develop SIGNATUREs for the
PROTEIN KEYS.
• These SIGNATUREs must be used to
CLASSIFY the proteins correctly into
their respective SUPERFAMILIES.
• Performance and Speed are important
Signature for Keys
• Accurately/concisely represent the keys.
• Signatures can be simple statistics like
mean, median etc. of the keys or a complex
combination of features.
• What ever may be the choice of
Signature/s, it/they must be able to
perform extremely fast and accurate
classification of the protein/s.
Choice of Classifier/Tool
• Criteria:
1. ACCURACY
2. SPEED
Final Product
A software that takes in “keys” as input
and classifies it correctly.
There must be a check if the “new”
protein-keys already exists in the
system.