|
AMITAVA BHATTACHARYYA
Email : abhattacharyya@amity.edu
bhattacharyya.amitava.in@gmail.com
|
Present : A1-A, Ground Floor, Single Storey,
Vijay Nagar, Delhi India- 110009
|
Objective:
To obtain a challenging position that utilizes my Bioinformatics/Computational skills. I am seeking an opportunity to be a significant contributor either as an individual or as a member of a team.
AREA OF INTEREST:
Bioinformatics: Protein Structure Prediction, evolutionary bioinformatics (comparative genomics) analyses , gene regulation, phylogeny/adaptive evolution, protein domain analysis and functional annotation, microarray bioinformatics, alternative splicing, proteomics, and protein structure analysis.
Computer science: Genetic Algorithm, HMM, Simulated-Annealing, Neural Network, Programming in different Languages, Algorithm Development.
Academic Experience:
Presently teaching at A.I.B., Amity University, Noida, U.P. as Adhoc Lecturer Bioinformatics (from 28th Jan 2008 onwards)
Responsibilities:
- Teaching—Bioinformatics Algorithms, Computational biology, Computational Genomics, Microarray, Biojava to B.Tech. And M.Sc. Bioinformatics and Biotechnology students.
- Student projects guided—
- Recognition of antigenecity in various rice genomes using Profile HMM
- Docking studies of HIV protease inhibitors
- Recognition of Alternative splice exons using Paired HMM
- Identification of Alternative spliced proteins using Genetic Algorithms
Taught “Machine Learning” and “Computer Graphics” at A.P.S. university, Rewa, M.P. (from Dec 2007 to Jan 2008)
EDUCATION
MASTER’S PROJECTS
-
M.Sc. Mini Project
Theoretical understanding of the protein folding problem using the Dill’s HP reduced model and Genetic algorithm (MATLAB Programming based )
One of the most important problems in biology is to establish or predict the secondary and tertiary structure adopted by a folded protein or peptide molecule from knowledge of its primary structure. Due to the size of most proteins, full atomistic simulations are impossible, giving rise to the need for reduced models, in which beads replace groupings of atoms. These models can be used to capture the underlying physics of the folding process. The reduced model used for the project is the Dill’s HP model, where the 20 amino acids are reduced to 2 hydrophobic and polar residues. This profoundly reduces the complexity of the problem understanding. Genetic algorithms are a search technique based on natural selection, where the best or fittest individuals, representing solutions to the search, in a population survive from one generation. Individuals are subjected to genetic operators analogous to mating and mutation in natural systems. The folded state of proteins lies at or near the protein's global minimum conformational energy. The deduction of the folded state can therefore be formulated as a search through the conformational space of the protein to find the global minimum energy. The programming platform that has been used in the project is Matlab.
The results were obtained for a sequence of length 50 and the optimization was carried for 100 generations. Better optimization results can be obtained on running the Genetic Algorithm on faster machines and for more number of generations. The experiment was carried out for a sequence of length 50 and was carried for 100 generations. The results show that the genetic algorithm is an efficient search tool, especially for large conformational spaces that pose a computational challenge. The optimized structures of the protein obtained on a lattice plane, provides a suggestive two dimensional structure of the protein. This is a useful feature since this method although crude works quickly and is able to provide a rough idea about how an amino acid sequence might fold into. This becomes even more useful in the case of new protein sequences. |
M.Sc. Major Project
Analysis Of Fourier Transform images for all protein conformations in 2-D HP lattice space
Folding of the amino acid chain into the functional structure of the protein has intrigued scientists over the many decades. Many methods have been proposed in this regard. All processes although successful to a certain extent have their limitations. Spectral analysis techniques are being used in a various domains of biological research. This project was an attempt to consolidate the information in the analysis of the conformational space of the protein using the Fourier transform image analysis technique. Fourier transform was applied on every sequence of the conformational space of protein sequences. Due to limitations of the computational power the Dill’s HP reduced model had been used.
Most tools for biomolecular sequence analysis have traditionally been computer-aided character string processing algorithms, in which the results are blindly presented as the processing output, without providing sufficient intuitive understanding of their meaning. In contrast, the Fourier images used here do not even search for obvious clues in the “character string domain” and thus, can be used to complement the existing tools that directly look into structure patterns, satisfying the need for direct visualization of various properties of biomolecular sequence structures. An important advantage of these visual tools is their flexibility. Once a pattern appears to exist, we have the opportunity to modify the values of these parameters in ways that will enhance the appearance of these patterns, thus clarifying the significance. Close observation of the Fourier transform images provides idea about the energy of the structure. The images show that blurring is proportional to the energy. The lower energy structures show prominent bright spots and minimum blurring. As the energy increases the blurring of the images also increases. The blurring or the haze is formed due to the increase in the number if the secondary maxima. Thus it can easily be deduced that that the compactness of the structure is directly proportional to the number of primary maxima. The secondary maxima increase as the structure looses its compactness. It was inferred that the correlation of the image pattern to the structure shows a relation with the energy score. Analysis of the whole conformational space presents before us the fact that most of the structures are energetically irrelevant. Only a small part of the total space can be considered worth considering. It is from this small portion the functional conformations originate. A consideration of these structures can provide greater information about the folding patterns and other functional aspects of the real proteins.
|
KEY SKILLS:
TECHNICAL SKILLS
-
Platforms : Windows 98 / Xp / NT / 2000, Linux(Fredora and Red Hat)
Software : MS Office Pro 97/2000/2003, Various Bioinformatics Softwares- CADD,
modeling, simulation and structure prediction
Programming : C, C++, Perl, Matlab, Visual Basic, Java core,
J2EE (Servelet, Socket, JSP, RMI, Swing)
Database : SQL, Pl-SQL, Oracle8i, JDBC, Microsoft Access
Web Programing : HTML, XML, Java script
-
Bioinformatics Skill:
Sequence Analysis: Being student of bioinformatics I have good understanding of algorithms used for alignment of two or more sequences. Developed Matlab programs for Dot Matrix sequence comparison and alignment of two sequences using Dynamic programming algorithm.
Databases: familiar with computer storage of sequences and sequence formats. Have good understanding of different databases for sequence and structure, like NCBI, EMBL, SWISS PROT, PIR sequence databases and PDB, MMDB, SCOP, CATH structure databases.
Database Searching: have working experience with different database search program like NCBI’s BLAST, ENTREZ. Also have good concept of BLAST and FASTA search algorithms.
Phylogenetic Analysis and clustering procedures: understand concept of character based (Parsimony, Compatibility) and distance based (UPGMA, NJ) methods for phylogenetic analysis.
Protein Structure Prediction: protein secondary structure prediction and very well familiar with Homology modeling, ab-initio method of structure prediction, Threading techniques etc.
Computer Algorithms: in-depth knowledge and thorough understanding of Genetic Algorithm, Dynamic Programming Algorithms, HMM, Neural network, etc.
Software Knowledge: Molecular Dynamics – VMD, TINKER, GROMACS
Visualization – Cn3D, RASMOL, SPDBV
Structural Editing – CHIMERA
Receptor-Ligand Docking – Dock6, Schrödinger
Phylogeny package – Phylip, Argus lab
Homology modeling software -- Modeller
SEMINAR ATTENDED:
- Indo-Russian International seminar on Recent trends in Applied Bioinformatics and Medi-informatics organized by IIIT (Indian Institute of Information Technology), Allahabad during Nov. 6-7th, 2005.
- National workshop cum training programme on ‘BIO-COMPUTING’ organized by IVRI (Indian Veternary Research Institute), Izatnagar, Bareilly during Mar. 24-27th, 2007.
Interpersional Skill:
Leadership, good communication skills, innovative, research aptitude , problem solving, optimistic, hard-working
Linguistics:
English
Bengali
Hindi
|
|
