Annotation Transfer for Genomics

1
Annotation Transfer for Genomics Quantification
of the Relation Between Protein Sequence
Structure and Function Wilson et al 2000 J. Mol.
Biol. 297,233-249
2
INTRODUCTION
Question Can we measure quantitavely the degree
to which Structural information Functional
information can be transferred between pairs of
proteins sequences at various levels of
similarities If yes we can infer structure or
function
3
INTRODUCTION
The problem of genome annotation genome
analysis-gtfunctional annotation of all the gene
products. From all known sequence only few of
them function have been determined
experimentally. homologues of newly identified
genes can be identified via database search
based on the fact that sequence similarity
implies structural and functional
similarities. However structural and
functional annotation should be transferred
with caution
4
PAST RESULTS
Sequence structure
Structural divergence when expressed by a rms
separation of matching alpha carbon atoms is an
exponentional function of sequence
divergence. The reliability of structural
annotation depends on the Sequence identity of
the homologous proteins Twilight zone ISS
5
PAST RESULTS
Sequence-function Gene function generally
based on sequence similarity. Pattern of hits
among different phylogenic groups focus on key
motifs associated with function
6
PAST RESULTS
Sequence-structure-function predict the
structure of an uncharacterized
sequence predict the function based on the
limited repertoire of function known to occur
with that structure. example fetrow et al.
1998. Prediction of structural profiles based
on threading ab initio methods and then search
with these against a known structure in order
to predict function.
7
PAST RESULTS
Sequence-structure-function Russel et al.
1998 discuss identification of structural
binding sites in predicting function. Hegyi
Gerstein 1999 investigated in which folds were
associated with functions. Most folds are
associated with one or two function. Excepti
on of a few folds such as TIM barrel that
could carry out numerous function.
8
Wilson et al. analysis
Sequence-structure-function Wilson et al 2000
strategy Pairwise comparisons of protein
sequence, structure and function among proteins
that share the same fold. Assessment of trends
relating sequence, structure and function consi
der the implications for structural and
functional annotation transfer
9
Alignment of 30,000 Pairs from SCOP
A) The basic unit of comparison a pair of
protein domains Identification of protein
domain in SCOP database B) Selection of
pairs There are 3.9 million possible pairs in
SCOP database Selection of 30,000 pairs
similarities from the total of SCOP at
different level of similarities All the
selected pairs where 50 residues in length and
were drawn from the four major SCOP
secondary structural classes all a, all b,
a/b,ab.
10
Alignment of 30,000 Pairs from SCOP

• B) Selection of pairs
• Same fold different function
• Different fold different function
• Same fold same function

11
Alignment of 30,000 Pairs from SCOP
B) Selection of pairs at different Level of
the SCOP database fold superfamily fami
ly
12
RMS and percent identity
RMS as a function of percentage identity
R0.21e 0.0132H RRMS H100-I I
identity Chothia and lesk (1986)
13
Structure sequence relationship
In the twilight zone structural similarity is
more significant than sequence similarity.
14
Structure sequence relationship
For pairs more than 30 identity a given pair
has more significant sequence similarity
than structural similarity. Because structure
is always more conserved than sequence, so
usually a given amount of structural similarity
is not as significant as a corresponding amount
of sequence similarity.
15
Structure sequence relationship
For closely related sequences, difference in
sequence similarities are more meaningful For
highly divergent sequences that share the same
Fold, the difference in structural similarities
are more Significant.
16
Linking sequence and structure to function
Difficulties of functional comparison there is
a well characterized relationship between
sequence and structure But the relationship
between structure to function has not been yet
characterized very well.
17
Linking sequence and structure to function
Difficulties of functional comparison The main
obstacles are absence of clear measure of
functional similarity. definition of
function is very often vague. So there are no
functional equivalent to the hierarchical fold
classification for domains in the PDB.
18
Functional Classification of pairs
New classification of pairs Pair can have no
functional similarity Pair can have three
levels of similarities
19
Functional Classification of pairs
New classification of pairs Pair can have
three levels of similarities (i) General
similarity Both domains are enzymes or no
enzymes (ii) Same functional class Both domains
share the three component of their EC enzyme
number 1.1.1.1 alcohol dehydrogenase 1.3.1.1
cortisone beta reductase (iii) Same precise
function Share three components of their
EC enzyme number. 1.1.1.1. Alcohol
dehydrogenase 1.1.1.3. Homoserine dehydrogenase
20
Functional Classification of pairs
Divide the pairs by general function enzyme
paired with non enzymesno general
function enzyme paired with enzyme same
general function non-enzymes paired with non
enzymes (same general function)
21
Functional Classification of pairs
EC System
Class Reaction nature Subclass Chemical group
On which enzyme act Sub subclass Acceptor
group Individual enzyme
22
Structure and Function
Similarity in structure and function is less
straightforward Than that between similarity in
sequence and function.
-X- non-enzyme with same functional
class -D- enzyme with same functional
class --X-- non-enzyme with same precise
function --D-- enzymes with the same precise
function
23
Structure and Function
Similarity in structure and function is less
straightforward Than that between similarity in
sequence and function.
-X- non-enzyme with same functional
class -D- enzyme with same functional
class --X-- non-enzyme with same precise
function --D-- enzymes with the same precise
function
24
Discussion conclusion
Assessment of the transfer of functional and
structural annotation by analyzing the
relationship between similarity In sequence
structure and function has been made. 30,000
pairs shows quantitative sequence structure
relationship As mentioned in previous
work. Measure of functional similarity provides
curves relating Function to sequence and
structure. When relating to functional
conservation to sequence Divergence, we find
distinct thresholds at 40 for precise Function
and 25 for functional class. Percent of
Identity is a more simple and direct measure of
functional divergence than other scoring
system
25
Discussion conclusion
Assessment of the transfer of functional and
structural annotation by analyzing the
relationship between similarity In sequence
structure and function has been made. 30,000
pairs shows quantitative sequence structure
relationship As mentioned in previous
work. Measure of functional similarity provides
curves relating Function to sequence and
structure. When relating to functional
conservation to sequence Divergence, we find
distinct thresholds at 40 for precise Function
and 25 for functional class. Percent of
Identity is a more simple and direct measure of
functional divergence than other scoring
system
26
Discussion conclusion