Project titles for:
Project titles for: 2007-09
- Comparison of software for Analysis of Molecular Marker Data
- In silico Prediction of miRNA in Rice Genome
- Web Tools for Bioinformatics
- In silico Structure Prediction of enzyme Arsenite methyltransferase
- Comparison of various software for QTL Analysis
- QSAR studies of Tuberculosis Inhibitors
- Determination of common transcription factor binding site in promoter region of abiotic stress resistance gene in rice
- SNP Mining in Rice Genome
- Protein Modelling of Betaine Aldehyde Dehydrogenase-2 in Rice
- Data Mining of Chemical Substructures for Biological Efficacy
- Plant Disease Database mined from PubMed
- Comparison of Protein Structure Prediction Methods
- Virtual High Throughput Screening for Influenza Virus Inhibitors
- Combinatorial Libraries for Screening against Tuberculosis Inhibitors
Two proteins (1FNA and 2PQT) with their structural information known were obtained from RCSB Protein Data Bank along with their amino acid sequences in fasta format. Using molecular viewers CHIMERA and DEEPVIEWER and other tools all the information regarding these proteins was recorded and tabulated. Then fasta sequences of these proteins were submitted to the secondary structure prediction servers. There were many servers but only three servers (GOR V, PSI-PRED, YASSP) were used. These servers predict secondary structures using different approaches (PSI-PRED = neural network, GORV = GOR, YASSP = SVM). When servers returned their predictions then these predictions were compared with the original data. By comparing the prediction results with the original data it was concluded that server GOR V with GOR Method had predicted the secondary structure with maximum accuracy, so it was concluded that this server is better for secondary structure prediction. Similar protocol was followed for tertiary structure prediction servers. Both the proteins were submitted to servers 3D-JIGSAW and CPHMODEL of homology approach, PHYRE and SAM-T08 of threading approach and I-TASSER serves of ab initio approach. From comparisons of predictions returned by these servers with the original data it was concluded that server CPHMODEL of homology approach, server SAM-T08 of threading approach and I-TASSER server of ab initio approach had predicted the tertiary structure with greater accuracy then other servers. These servers are recommended to be better servers than others. Since the structure of these proteins were already known so homology based servers have withdrawn the structural information about these proteins from their own structures. In case the structure of proteins are not known then it would be a real check, even in that cases the proteins having greater amino acid sequence identity are predicted with greater accuracy. 19 It was also concluded that if a protein is having good amino acid sequence similarity (80-90 percent) then it is always better to choose homology approach specially when the protein is of greater length (> 300 AA) because for this ab initio approaches takes so much time and fold recognition approaches returns the same proteins as that of homology. If the proteins are of small length and having very poor sequence identity with the already known proteins then ab initio approaches always preferable.