<s>
The	O
PageRank	B-Algorithm
algorithm	I-Algorithm
has	O
several	O
applications	O
in	O
biochemistry	O
.	O
</s>
<s>
(	O
"	O
PageRank	B-Algorithm
"	O
is	O
an	O
algorithm	O
used	O
in	O
Google	B-Application
Search	I-Application
for	O
ranking	O
websites	O
in	O
their	O
results	O
,	O
but	O
it	O
has	O
been	O
adopted	O
for	O
other	O
purposes	O
also	O
.	O
</s>
<s>
According	O
to	O
Google	B-Application
,	O
PageRank	B-Algorithm
works	O
by	O
"	O
counting	O
the	O
number	O
and	O
quality	O
of	O
links	O
to	O
a	O
page	O
to	O
determine	O
a	O
rough	O
estimate	O
of	O
how	O
important	O
the	O
website	O
is	O
,	O
"	O
the	O
underlying	O
assumption	O
being	O
that	O
more	O
important	O
websites	O
are	O
likely	O
to	O
receive	O
more	O
links	O
from	O
other	O
websites	O
.	O
)	O
</s>
<s>
The	O
relative	O
importance-measuring	O
property	O
of	O
the	O
PageRank	B-Algorithm
link	O
analysis	O
algorithm	O
could	O
be	O
used	O
to	O
identify	O
new	O
possible	O
drug	O
targets	O
in	O
proteins	O
.	O
</s>
<s>
A	O
PageRank-based	O
algorithm	O
could	O
identify	O
important	O
protein	O
targets	O
in	O
the	O
pathogen	O
organism	O
better	O
than	O
a	O
method	O
considering	O
only	O
the	O
number	O
of	O
incoming	O
edges	O
(	O
in-degree	O
)	O
of	O
a	O
node	O
in	O
the	O
metabolic	O
network	O
.	O
</s>
<s>
By	O
using	O
the	O
available	O
metabolic	O
networks	O
,	O
it	O
is	O
possible	O
to	O
find	O
important	O
nodes	O
with	O
link	O
analysis	O
algorithms	O
,	O
like	O
PageRank	B-Algorithm
.	O
</s>
<s>
A	O
PageRank-based	O
scoring	O
method	O
could	O
detect	O
important	O
nodes	O
that	O
are	O
not	O
hubs	O
and	O
therefore	O
might	O
be	O
better	O
drug	O
targets	O
.	O
</s>
<s>
The	O
PageRank	B-Algorithm
of	O
a	O
node	O
A	O
is	O
the	O
stationary	O
limit	O
probability	O
distribution	O
that	O
the	O
random	O
walker	O
is	O
at	O
node	O
A	O
.	O
</s>
<s>
This	O
personalized	O
PageRank	B-Algorithm
may	O
identify	O
other	O
related	O
proteins	O
to	O
the	O
disease	O
.	O
</s>
<s>
However	O
,	O
by	O
using	O
only	O
the	O
personalized	O
PageRank	B-Algorithm
to	O
identify	O
important	O
nodes	O
,	O
hubs	O
still	O
get	O
a	O
high	O
score	O
on	O
average	O
.	O
</s>
<s>
To	O
find	O
non-hub	O
important	O
nodes	O
instead	O
,	O
we	O
should	O
consider	O
scoring	O
the	O
nodes	O
by	O
their	O
"	O
relativized	O
personalized	O
PageRank	B-Algorithm
"	O
;	O
i.e.	O
</s>
<s>
their	O
personalized	O
PageRank	B-Algorithm
scores	O
over	O
the	O
number	O
of	O
edges	O
pointing	O
towards	O
them	O
(	O
over	O
their	O
in-degree	O
)	O
:	O
</s>
<s>
The	O
relativized	O
personalized	O
PageRank	B-Algorithm
(rPPR(v )	O
)	O
for	O
a	O
node	O
v	O
is	O
given	O
by	O
:	O
</s>
<s>
where	O
PpageRank(v )	O
is	O
the	O
personalized	O
PageRank	B-Algorithm
score	O
of	O
node	O
v	O
,	O
and	O
d_(v )	O
is	O
its	O
in-degree	O
.	O
</s>
