DR. NARAYAN BEHERA

Sequence alignments of proteins and RNAs by genetic algorithm

Computational methods have the power to infer biological information from DNA sequence alone. Sequence alignment provides knowledge on structural, functional and evolutionary relationship about bio-molecules. It is necessary to find optimal or best possible alignment to discover this information. An alignment algorithm finds an optimal alignment for two sequences (given a scoring system). A scoring matrix makes a statement about the probability of observing residue pairs in real alignments. Dynamic programming generates an alignment by comparing every pair of characters in the two sequences. It finds an optimal alignment for a given additive alignment score. Multiple sequence alignment (msa) provides knowledge about the conserved regions in protein families. It also helps to determine the evolutionary relationship of species. Given a set of homologous sequences, multiple alignments are used for predicting the secondary or tertiary structure of new sequences (such as RNA or proteins). Genetic algorithm is a powerful method to obtain information about multiple sequence alignment. This algorithm makes use of the principles of mutation, selection and recombination to solve an optimization problem. Here the alignment score needs to be optimized. A scoring function for multiple alignments can be written as the sum of all pair-wise alignments. The optimal alignment is based on obtaining the best score.

The genetic algorithm is an important method to study RNA alignment while taking into account the primary sequence and secondary structure. Accurate alignment provides important non-experimental information to gather reliable knowledge about the RNA secondary structure. Our understanding of the folding process in vivo is incomplete. However, homology analysis based on sequence alignments does not have these types of limitations. The folding of very long RNA sequences generally has large search spaces. The correct solution of this problem requires long computational time. One is forced to approximate a best solution. This genetic algorithm approach offers a unique advantage in this regard. An important part of the present work focuses on the design of better mutation operators that improves the efficiency and accuracy of the genetic algorithm. This involves incorporation of new objective functions that make use of more complex gap insertions as well as crossover operations. This has the potential to provide better accuracy in multiple sequence alignments and hence prediction of the structures of RNAs and proteins.

Micro-array data analysis by evolutionary computation

Micro-array data and pathway mapping for certain inherited metabolic disorders are being analysed. A database is being developed for the analysed results and expression patterns of the pathways. This database can also include the micro array data generated by different researchers throughout the world. The basic idea is to analyse the available micro-array expression data and then map the expression patterns of genes in the affected metabolic pathway. The comparative analysis of data might lead to the identification of the newer set of genes affected during the disease process. One can also keep track of the variability of the expression pattern of the known/new gene sets across differ rent types of data. The changes in the gene expression levels are relatively minor in such metabolic syndrome. Therefore, the data analysis requires a very sensitive method. A new and very sensitive clustering algorithm is being developed that can identify even very minute changes in the gene expression patterns. Algorithms that make use of evolutionary computation method are being developed to address the problem of refined micro-array data analysis.

Selected publication list:

• Genus-two Correlators of Critical Ising Model (1989) Physical Review D 40, 1993-200 (with Malik R P & Kaul R K)
  
• Level-one SU(3) Wess-Zumino Model on Higher Genus Riemann Surfaces (1990) Physical Review D 41,478 – 483 (with Malik R P & Kaul R K)
  
• Correlators of Conformal Field Theories from Their Characters (1990) Modern Physics Letters A Vol. 5, 2643- 2649 (with Malik R P & Kaul R K)
  
• Level-one and two SU(2) Wess-Zumino model on Higher Genus Riemann Surfaces (1991) Physical Review D43, 1243-1253(with Malik R P & Kaul R K)
  
• An Investigation into the role of Phenotypic Plasticity in Evolution (1995) Journal of Theoretical Biology Vol.172, No. 3, 225-234 (with Nanjundiah V)
  
• A Compact Proof of Fisher's Fundamental Theorem for Multiple Loci (1995) Journal of Genetics Vol. 74, Nos.1&2, 19-24
  
• Variational Principles in Evolution (1996) Bulletin of Mathematical Biology Vol.54, 175-202.
  
• The Consequence of phenotypic plasticity in cyclically varying environments: a genetic algorithm study (1996) Journal of Theoretical biology, Vol.178, No.2, 135-144 (with Nanjundiah V)
  
• trans-Gene Regulation in Adaptive Evolution: a Genetic Algorithm Model (1997) Journal of Theoretical Biology Vol. 188, 153-162(with Nanjundiah V)
  
• Phenotypic diversity and stability in ecosystem processes (1999) TheoreticalPopulation Biology 56, 29-47(with Loreau M)
  
• The influence of life history differences on the evolution of reaction norms (2002) Evolutionary Ecology Research, Vol. 4, 1-25 (with de Jong G)
  
• Phenotypic plasticity can potentiate rapid evolutionary change (2004) Journal of Theoretical Biology, 226, 177-184 (with Nanjundiah V )
  
• Evolution of mutualism through spatial effect (2004) Journal of Theoretical Biology, 226, 421-428 (with Yamamura, Y, Higashi M & Wakano J Y)
  

Research papers accepted, submitted and/or to be submitted

• Increase of species coexistence due to spatial effect (2006) (with Yamamura N & Higashi, M) (submitted)
  
• Evolution of niche width in haploid population (2006) (with de Jong G) (submitted)
• Multiple sequence alignment by evolutionary computation (2006) (Goyal, P., Chanchal, K., Flavia, D., Lokanath, K., Prajkta, K., Siva, T. & Tanushree, B (to be submitted)

Funding received

• Department of Information Technology, New Delhi
• Novel sequence alignment of proteins and RNAs.

• Philips Research Asia, Bangalore
• Data mining algorithm to analyze Micro-array gene expression data.