You are on page 1of 3

ROLL NUMBERS: BSCSF12M054

BSCSF12MO11
NAMES:
RAB

MALIK-UBAID-ULADEEL-UL-

RHAMAN
Overview Of The Given
Assignment:
Intro:
In this assignment an algorithm is discussed relevant to issues
that are important according to biological point of view. This
algorithm finds the protein complexes from large interaction
networks . By knowing the protein complexes we can predict the
function unknown function protein ,new function for function
known protein and can help us to understand the certain
biological issues.
Summary :
In a network a protein represents a node and an edge is
represented by interaction This algorithm is of polynomial time
complexity O(N^3) ,where N is the number node implemented
by java programming.
The proposed algorithm is based on following four parameters:
1) Density of the cluster(protein complex)

d= |E|/|E|(max) =2*|E|/|N|*(|N| -1)


The value of density should be equal to the to our desired
value .
It tackles every cluster in the graph and removes them one
by one until there is
a single edge left in remaining graph.
2) Weight of an edge : It is the common number of neighbor
between two nodes. Time complexity of finding weight of
graph is O(N^3) which can be reduced to O(|N|||E|).
3) Weight of a node: It is the sum of weights of edges
connected to the node.
Time complexity of weight of nodes is O(|N^2|).
4) CP(cluster property): If a node is the part of the cluster it
should be connected to reasonable number of edges present
in cluster
i.e d*(|N|-1) edges in cluster.
CP is found by formula : |EC|/d*(|N|) where EC is the number
edges between the node and the cluster. We can threshold
value of CP in the range 0<CP<=1.
The weight of node is used to generate and sort the neighbors.
Before we add a node to a cluster we see priority of that node
which based on two measures
(1)weight of the node between a neighbor and a cluster.
(2)weight of edge of a neighbor and a cluster.
Worst case time complexity for this job is O(|N|*CS),Where CS is
the size of cluster.
RESULTS:
This algorithm have been implemented on a network consisting of
287 interactions and 300 proteins , where values of d=0.60 and
CP= 0.50 were given. The algorithm generated 1and size four and
density more than 0.6and size three and density 1.

You might also like