Diapositive 1 - Institut Pasteur

Download Report

Transcript Diapositive 1 - Institut Pasteur

Mutivariate statistical Analysis
methods
Ahmed Rebaï
Centre of Biotechnology of Sfax
[email protected]
Basic statistical concepts and
tools
Statistics
 Statistics
are concerned with the
‘optimal’ methods of analyzing data
generated from some chance
mechanism (random phenomena).
 ‘Optimal’ means appropriate choice
of what is to be computed from the
data to carry out statistical analysis
Random variables
 A random
variable is a numerical
quantity that in some experiment, that
involve some degree of randomness,
takes one value from some set of
possible values
 The probability distribution is a set of
values that this random variable takes
together with their associated
probability
The Normal distribution
Proposed by Gauss (1777-1855) : the distribution
of errors in astronomical observations (error
function)
 Arises in many biological processes,
 Limiting distribution of all random variables for
a large number of observations.
 Whenever you have a natural phenomemon
which is the result of many contributiong factor
each having a small contribution you have a
Normal

The Quincunx
Bell-shaped
distribution
Distribution function

The distribution function is defined F(x)=Pr(X<x)
t
F ( t )   f ( x )dx where




1
f(x)
e
2²
( x  )²
2 ²
F is called the cumulative distribution function (cdf)
and f the probability distrbution function (pdf) of X
 and ² are respectively the mean and the variance of
the distribution
Moments of a distribution

The kth moment is defined as

'k  E( x )   x f ( x ) dx
k
k



The first moment is the mean
The kth moment about the mean  is

k  E( x   )   ( x   ) f ( x ) dx
k
k


The second moment about the mean is called
the variance ²
Kurtosis: a useful moments’ function

Kurtosis
4=4-3²2
4  0 for a normal distribution so it
is a measure of Normality

Observations


Observations xi are realizations of a
random variable X
The pdf of X can be visualized by a
histogram: a graphics showing the
frequency of observations in classes
Estimating moments

The Mean of X is estimated from a set of
n observations (x1, x2, ..xn) as
1 n
x   xi
n i 1

The variance is estimated by
n
1
2
2
( xi  x )
Var(X) =  

n  1 i1
The fundamental of statistics
Drawing conclusions about a
population on the basis on a set of
measurments or observations on a
sample from that population
 Descriptive: get some conclusions
based on some summary measures
and graphics (Data Driven)
 Inferential: test hypotheses we have
in mind befor collecting the data
(Hypothesis driven).

What about having many variables?



Let X=(X1, X2, ..Xp) be a set of p variables
What is the marginal distribution of
each of the variables Xi and what is
their joint distribution
If f(X1, X2, ..Xp) is the joint pdf then the
marginal pdf is
f ( X i )   f ( x1 ,.., xi 1 , xi 1 ...,x p )dx1 ....dx p
Independance

Variables are said to be independent if
f(X1, X2, ..Xp)= f(X1) . f(X2)…. f(Xp)
Covariance and correlation

Covariance is the joint first moment of
two variables, that is
Cov(X,Y)=E(X-X)(Y- Y)=E(XY)-E(X)E(Y)

Correlation: a standardized covariance
( X ,Y ) 

Cov( X ,Y )
Var( X ).Var( Y )
 is a number between -1 and +1
For example: a bivariate Normal

Two variables X and Y have a
bivariate Normal if
f ( x, y ) 

1
21 2 1   2

e
( x 1 )( y  2 ) ( y   2 )² 
1  ( x 1 )²

2




2

²


² 2 
1  
1
1 2
 is the correlation between X and Y
Uncorrelatedness and independence

If =0 (Cov(X,Y)=0) we say that the
variables are uncorrelated

Two uncorrelated variables are
independent if and only if their joint
distribution is bivariate Normal

Two independent variables are
necessarily uncorrelated
Bivariate Normal
f ( x, y) 


1
21 2 1   2
e
( x  1 )( y   2 ) ( y   2 )² 
1  ( x  1 )²

2




 1 2
 ²2 
1  2   ²1
If =0 then
f ( x, y) 
1
21 ²
e
 ( x  1 )² 


  ²1 
1
2 2 ²
e
 ( y   2 )² 


  ²2 

So
f(x,y)=f(x).f(y)

the two variables are thus independent
Many variables

We can calculate the Covariance or
correlation matrix of (X1, X2, ..Xp)
 v( x1 ) c(x1,x2 ) ........c(x1,x p ) 


 c(x1,x2 ) v( x2 ) ........c(x2 ,x p ) 




 c(x ,x ) c(x ,x )..........v( x ) 
2 p
p 
 1 p

C=Var(X)=

A square (pxp) and symmetric matrix
A Short Excursion into Matrix
Algebra
What is a matrix?
Operations on matrices
Transpose
Properties
Some important properties
Other particular operations
Eigenvalues and Eigenvectors
Singular value decomposition
Multivariate Data
Multivariate Data
 Data
for which each observation
consists of values for more than
one variables;
 For example: each observation
is a measure of the expression
level of a gene i in a tissue j
 Usually displayed as a data
matrix
Biological profile data
The data matrix
 x11 x12 ....x1 p 


 x 21 x 22 ....x 2 p 




 x x ....x 
np 
 n1 n 2
n observations (rows) for
p variables (columns)
an nxp matrix
Contingency tables


When observations on two categorial
variables are cross-classified.
Entries in each cell are the number of
individuals with the correponding
combination of variable values
Eyes colour
Hair colour
Fair
Red
Medium
Dark
Blue
326
38
241
110
Medium
343
84
909
412
Dark
98
48
403
681
Light
688
116
584
188
Mutivariate data analysis
Exploratory Data Analysis


Data analysis that emphasizes the use of
informal graphical procedures not based
on prior assumptions about the structure
of the data or on formal models for the
data
Data= smooth + rough where the smooth
is the underlying regularity or pattern in
the data. The objective of EDA is to
separate the smooth from the rough with
minimal use of formal mathematics or
statistics methods
Reduce dimensionality without loosing
much information
Overview on the techiques
 Factor
analysis
 Principal components analysis
 Correspondance analysis
 Discriminant analysis
 Cluster analysis
Factor analysis
A procedure that postulates
that the correlations between a
set of p observed variables arise
from the relationship of these
variables to a small number k of
underlying, unobservable, latent
variables, usually known as
common factors where k<p
Principal components analysis
A
procedure that transforms a set
of variables into new ones that are
uncorrelated and account for a
decreasing proportions of the
variance in the data
 The new variables, named principal
components (PC), are linear
combinations of the original
variables
PCA
 If
the few first PCs account for
a large percentage of the
variance (say >70%) then we
can display the data in a
graphics that depicts quite well
the original observations
Example
Correspondance Analysis
A method for displaying relationships
between categorial variables in a
scatter plot
 The new factors are combinations of
rows and columns
 A small number of these derived
coordinate values (usually two) are
then used to allow the table to be
displayed graphically

Example: analysis of codon usage and
gene expression in E. coli (McInerny, 1997)
A gene can be represented by a 59dimensional vector (universal code)
A genome consists of hundreds (thousands)
of these genes
Variation in the variables (RSCU values)
might be governed by only a small number
of factors
For each gene and each codon i calculate
RCSU=# observed codon /#expected codon
Codon usage in bacterial genomes
Evidence that all synonymous codons were not used with equal
frequency:
Fiers et al., 1975 A-protein gene of bacteriophage MS2, Nature 256, 273-278
UUU
Cys
UUC
Cys
UUA
Ter
UUG
Trp
CUU
Arg
CUC
Arg
CUA
Arg
CUG
Arg
Phe
0
Phe
3
Leu
*
Leu
12
Leu
7
Leu
6
Leu
6
Leu
3
AUU Ile
6
UCU Ser
5
UAU Tyr
4
UGU
10
UCC Ser
6
UAC Tyr
12
UGC
8
UCA Ser
8
UAA Ter
*
UGA
6
UCG Ser
10
UAG Ter
*
UGG
6
CCU Pro
5
CAU His
2
CGU
9
CCC Pro
5
CAC His
3
CGC
5
CCA Pro
4
CAA Gln
9
CGA
2
CCG Pro
3
CAG Gln
9
CGG
1
ACU Thr
11
AAU Asn
2
AGU
Multivariate reduction

Attempts to reduce a high-dimensional
space to a lower-dimensional one.
In other words, it tries to simplify the data set.
Many of the variables might co-vary, therefore there might
only be one, or a small few sources of variation in the
dataset
A gene can be represented by a 59-dimensional vector
(universal code)
A genome consists of hundreds (thousands) of these genes
Variation in the variables (RSCU values) might be governed by
only a small number of factors
Plot of the two most important axes
Lowly-expressed
genes
Highly
expressed
genes
Recently
acquired genes
Discriminant analysis



Techniques that aim to assess whether or
a not a set of variables distinguish or
discriminate between two or more groups
of individuals
Linear discriminant analysis (LDA): uses
linear functions (called canonical
discriminant functions) of variable giving
maximal separation between groups
(assumes tha covariance matrices within
the groups are the same)
if not use Quadratic Discriminant
analysis (QDA)
Example: Internal Exon prediction

Data:
 A set of exons and non-exons
 Variables : a set of features
 donor/acceptor
site recognizers
 octonucleotide preferences for
coding region
 octonucleotide preferences for
intron interiors
 on either side
LDA or QDA
Cluster analysis
A
set of methods (hierarchical
clustering, K-means clustering, ..)
for constructing sensible and
informative classification of an
initially unclassified set of data
 Can be used to cluster individuals
or variables
Example: Microarray data
Other Methods
Independant component analysis (ICA):
similar to PCA but components are
defined as independent and not only
uncorrelated; moreover they are not
orthogonal and uniquely defined
 Multidimensional Scaling (MDS): a
clustering technique that construct a
low-dimentional geometrical
representation of a distance matrix
(also Principal coordinates analysis)

Useful books: Data analysis
Useful book: R langage