1. Identify at least 5 web sites which applying data mining to solve

biololgy's problems. Also conclude the problem and the solution of

each web site in your own words.

http://gizmo.lbl.gov/DM_TOOLS/OPM/opm.html

Tools for defining, browsing and manipulating data in genomic databases using the Object Protocol Model.

http://strubix.com/

Structural bioinformatics technology resource building the database of dynamic 3D structures of protein-based drug targets.

http://bioinfo.ernet.in

The bioinformatics center at the university of Pune provides information in the area of biotechnology with an emphasis on virology, protein and nucleic acid sequences and structures, microbial strain data, and also access to related areas through networks.

http://www.genebio.com/

commercial institutions of the proteomics databases and software tools which are developed and maintained by the Swiss Institute of Bioinformatics (SIB). providing quality databases, software tools and services to the Life Science industry particularly in the field of proteomics.
 
 

http://www.informaxinc.com/

commercial developer of bioinformatics software for accelerated drug discovery, enabling to achieve greater insights into genetic function through biological data mining and integrated sequence analysis. There are some free trials.
 
 

2. What are the Artificial Neural Networks and EvolutionaryAlgorithms? Compare and identify their advantages and disadvantages.

Both Artificial Neural Network (ANN) and Evolutional Algorithms (EA)

are transformed from the general principle developed in nature.(refernce)

ANN

Model of architecture is structure of real neural network . The real neuron consists of synapses,axon and dendrite this allow the elctric signal to jump across from neuron to neuron.

Basic algorithm In ANN each neuron has its own unique threshold to weigh certain number of input and only the output over the treshold (output a 1)is fed to other neurons as input, for the neuron that the input is under the treshold (output a 0) and stays quiet. The value of treshold is set through the learning rules either by supervise(require teacher=training set) or unsupervise (not require teacher neuron produce their own output which is then further weigh)

Application ANN function as a pattern classifiers or pattern association in case that traditional technique do not work(complex input).

Disadvantage to ANN is that it is time cosuming and lack of rules to help in network construction.

EA

Model of architecture natural biological evolution (livings have adaptation to the environment and the fittest livings will be naturally selected to be survive)

Basic algorithm A large set of possible solutions for given problem is initially generated ( initial state in the problem is not already given ) then evaluate each solution to find the fit solutions and only the fit solutions can reproduce the other sets of solution until the output is close enough to the desired answer.

Application EA can give solution to any problem that has large search domain

Disadvantage If the problem has huge search space EA will generate an infinite number of solutions.