Biology Reference
In-Depth Information
The k × n matrix product E T H aggregates the mating structure for each
genotype. Without loss of generality, let E (1 ,i ) = E (2 ,i ) = ... = E ( t,i ) =1
and E ( t +1 ,i ) = E ( t +2 ,i ) = ... = E ( m,i ) =0. Then the i th row of E T H is
η ( v 1 )+ η ( v 2 )+ ... + η ( v t ). From the definition of a diversity graph we know
there are t/ 2 disjoint pairs, ( v p ,v q ), with p and q no greater than t , such that
η ( v p )+ η ( v q )= γ ( w i ). This means that the i th row of E T H is ( t/ 2) γ ( w i ).We
have just established the following result.
then E T H
Theorem
6.2. If (
V
,
W
,
E
,η,γ ) is
a
diversity
graph,
=
diag 2 E T e G .
The matrix equation in Theorem 6.2 succinctly separates the structure of the
graph, explained by E , from the labeling of the graph, explained by H and G .
Unfortunately, satisfying the matrix equation does not guarantee the graph is a
diversity graph because the aggregated information ignores the need of a mating
structure. As an example
11
1
1
111
1
E T H =(1111)
1
1
11
1
111
=(2)( 0000 )
= diag 1
2 E T e G.
This labeling of K 4 , 1 does not lead to a diversity graph since no pair of haplotypes
(no two rows of H ) add to form the single genotype (the row of G ).
We conclude this section with a discussion of a logical operator that helps
address the failure of Theorem 6.2 to characterize graphs with the stated matrix
equation. The logical join of a sequence of matrices is determined by the logical
operator “or” over each component of these matrices. The component-wise logical
join is defined so that 0
A 1 ,A 2 ,...,A s
}
is a logical decomposition of A if A is the logical join of the matrices in this set,
denoted:
0=0, 0
1=1,and1
1=1.Theset
{
A i = A 1
A 2
A s = A,
...
1 ≤i≤s
where we assume that all matrix elements are 0 or 1. For example, the matrices
on the left are a logical decomposition of the matrix on the right,
1100
1111 .
Such decompositions are used in the next section to characterize the graphs that
support diversity.
1010
1100
0101 = 1100
Search WWH ::




Custom Search