Clustering-coefficient- artificial intelligence, Computer Engineering

This programming assignment is about computing topological properties of Protein-Protein Interaction (PPI) networks. Recall that a PPI network is represented by a graph G=(V,E) where  nodes of V represent  proteins and an edge of E connecting  two nodes represents interacting proteins (either physically or functionally).  The properties include clustering coefficient and node centrality.  A detailed description of these properties  follows:

 Clustering coefficient

In many networks, if node u is connected to v, and v is connected to w, then it is highly probable that u also has a direct link to w. This phenomenon can be quantified using the clustering coefficient

Cu = 2nu/ k(k-1)

where nu is the number of edges connecting the ku  neighbors of node u to each other. In other words, Cu gives the number of 'triangles' that go through node i, whereas ku(ku -1)/2 is the total number of triangles that could pass through node u, should all of node u's neighbors be connected to each other.

The average clustering coefficient,<C>, characterizes the overall tendency of nodes to form clusters or groups. An important measure of the network's structure is the function C(h),which is defined as the average clustering coefficient of all nodes of degree h, i.e. with h adjacent vertices.

Closeness centrality of a node

Closeness-centrality is a measure of node centrality and uses information about the length of the shortest paths within a network; it uses the sum of the shortest distances of a node to all other nodes. The closeness-centrality of node u is defined as the reciprocal of this sum:

Cclo(u) = 1/(ΣvV dist (u , v)).

Implementation

Write 5 subroutines:

1.    Clustering-Coefficient

2.    All-Pairs-Shortest-Paths

3.    Create-Adjacency-Matrix

4.    Node-Centrality

5.    Get-Shortest-Paths 

The subroutine Clutering-Coefficient outputs 1) the average clustering coefficient <C>, 2) C(H), for all H values, and 3) the top 5 nodes (proteins) with highest cluster coefficient. The protein ids should be those used in the .sif file.

The procedure Create-Adjacency-Matrix takes as input argument the .sif representation of a graph and generates the adjacency matrix representation of the graph. It uses hashes (as defined in perl)  to map the protein ids into indexes of the adjacency matrix. 

The procedure All-Pairs-Shortest-Paths has as input argument the adjacency matrix representation of a graph. The subroutine returns a two-dimensional matrix, A, with A[i,j]  giving the length of the shortest path between nodes i and j.

 The subroutine Node-Centrality outputs the top 5 nodes according to the centrality measure, that is the 5 proteins that are most central. The protein ids should be those used in the .sif file.

 The subroutine Get-Shortest-Path takes in input a pair of nodes and the adjacency matrix representation of a graph and returns a shortest path  between the two nodes.  Call this subroutine from the main using as input parameters the two nodes with highest value of node-centrality.

 Data

In this assignment you analyze the Protein-Protein Interaction (PPI) graph of the herpes Kaposi virus. The file kshv.cys (available at t-square, Resources) contains such a graph in cytoscape format.

You select the subgraph of this network consisting of the nodes with degree equal or less than k,(k=7).

This will be the input to your perl program Properties-PPI-Networks.pl that calls the subroutines described above.

Electronically submit the following:

a.     the file Properties-PPI-Networks.pl containing the main program and all subroutines. 

b.    A document where you state the time complexity of each subroutine  as a function of the number of nodes and/or edges and explain how you obtained it.

 

Posted Date: 2/21/2013 4:57:21 AM | Location : United States







Related Discussions:- Clustering-coefficient- artificial intelligence, Assignment Help, Ask Question on Clustering-coefficient- artificial intelligence, Get Answer, Expert's Help, Clustering-coefficient- artificial intelligence Discussions

Write discussion on Clustering-coefficient- artificial intelligence
Your posts are moderated
Related Questions

Q. Write a pseudo code to find sum of two functions? Let's write a pseudo code to find sum of two functions f(A) + f(B). In first algorithm we will not use locking. Process

Define Linear Probing Linear Probing:  The easiest way to resolve a collision is to begin with the hash address and do a sequential search by the table for an empty location.

1. It is hard even for a highly skilled experts to abstract good situational assessment when he is under time pressure. 2. Expert systems perform well with specific t

Your professor wants you to fill a two-dimensional N by N matrix with some numbers by following a specific pattern. According to his explanation as in the figure below, you have to

Q Develop a program that accepts decimal number and converts it into given base. Fractional numbers are allowed and negative numbers are not allowed. Also check that entered number

Q. Illustrate LCD Technology? The technology behind LCD is known as Nematic Technology since the molecules of liquid crystals used are nematic which implies that rod-shaped. Th

What is the draw back of assigning one bit position to each control signals? Assigning individual bits to every control signal results in long microinstructions due to the numb

An expert systems is software that contains a knowledge base of facts and relationships and has the ability to make inferences based on that knowledge base. An experts sys

Search mechanisms in Prolog: Here we can needs this simple Prolog program to describe how Prolog searches as:president(X) :- first_name(X, georgedubya), second_name(X, bush).