Protein motifs with the secondary characteristics of hydrophobicity and polarity

• Related paper: Xie,J. and Kim, N.K., Bayesian Models and Markov Chain Monte
  Carlo Methods for Protein Motifs with the Secondary Characteristics, 2005,
  Journal of Computational Biology, Vol. 12, No. 7.
  

• Dataset 1glqA2 protein family used in the paper, sequences in FASTA format
  

• Multiple structural alignment of 1glqA2 family originally reported in CATH
  

• Programs of identifying protein motifs
  Our main development is the mixture model of protein motifs with both amino acid
  sequence and side chain polarity features. To search for a motif in multiple protein
  sequences, we suggest to first run the standard Gibbs sampling approach several
  times. If you think no conserved motif was identified, then run our mixture model.
  
  Download the programs
  
  
  • Standard Gibbs sampling approach: Unix (IBM AIX 4.3.3) executable
    
    You need to provide three arguments, the multiple sequence data set (FASTA
    format), motif width, and the number of repeated runs (each from different
    randome initial values). You may try number_of_runs = 3.
    
    Use the command: aa.out data_file motif_width number_of_runs
    
    
  • Our mixture model for both amino acid sequences and side chain polarity:
    Unix (IBM AIX 4.3.3) executable
    
    You need to provide three arguments, the multiple sequence data set (FASTA
    format), motif width, and the number of iterations. Some choices of the iteration
    number are 1000, 5000, and 10000.
    
    Use the command: mix.out data_file motif_width number_of_iterations