Project in Bioinformatics- Sequence Based Prediction- Research Report Assignment Help

Internal Code: 1HDGI

Research Report Assignment Help:

Task: Introduction Proteins are the products of translation of RNA molecule. They are the actual molecules that are taking part in interactions in the biological system. These interactions play an important role in the cellular cascades and pathways. Out of all the proteins identified till date only 30% of their structures have been determined. Methods for predicting protein interactions based on structure have been developed. However due to the lack of available structures much analysis cannot be depended on this method. Hence we propose to develop a protein sequence based method to predict interactions. Research question If two protein sequences consisting of 20 amino acid letters is given a method is to be developed to predict whether they interact or not. If they interact what is the probability of the tool to produce a true positive and false negative results? The tool is to be checked for the sensitivity and specificity of the results. Comparison with other existing software’s need to be done to ensure the accuracy of the predictions made. Aim/ Objective The aim of this work is to develop a web based server or a desktop application to predict the interactions between proteins based on their sequences only. Methods suggested to use on solving the research question The principle behind the development of a tool is co-evolution of proteins. In a pair of interacting proteins, if there is a change in the sequence of one protein, the other protein adjusts itself to accommodate that change over time. Hence there will be a connection between the interacting proteins. There are three criteria to be used while performing this method: 1. Classification of interaction and non-interacting partners on amino acids based on their physico- chemical properties such as hydrophilicity, hydrophobicity, ionic and non-ionic interactions. A chart is to be prepared based on which two amino acid pairs interact and not. 2. Position specific scoring matrix: A PSSM is to be constructed based on the distance of the residues in the sequences. The relative distances of amino acids also play an important role in the prediction. Based on this random sequences has to be prepared to create a positive and negative dataset for protein interaction prediction. The positive dataset includes all interacting pairs and negative includes non-interacting pairs. A Support Vector Machine is to be trained using the above data and appropriate kernels should be set. A set of true interactions and non interactions may be used as test data set based on the available literature. The tool after testing may be deployed as a web server or maintained as a desktop application. Influence to society? Why it is needed to do this study? Protein-protein interactions are the essence of any biological process. Defect in any of the numerous interactions in the body can lead to the development of disease. Since the structure of most of the proteins are not available with accuracy, a sequence based prediction method can help to identify interactions. The most important part of identifying the interactions is in the area of drug development. Existing drugs act by blocking or facilitating protein protein interactions. Based on the simple information of sequences it may be possible to develop drugs against a particular target disease. From sequence information we may be able to identify certain binding motifs which can be used to develop peptidomimetics or circular peptide drugs for blocking motif interactions. Hence the development of this tool can be well utilized by the scientific community for developing high quality drugs that can be delivered to the welfare of society.