CRE_conservativity {BALCONY}R Documentation

Calculate cumulative relative entropy score


This function calculates cumulative relative entropy score according to: Hannenhalli and Russell (2000).


CRE_conservativity(alignment, hmmbuild_path=NULL, pairwiseAlignemnt_scores=NULL)



An alignment object read with read.alignment function


(optional if running under UNIX) The aboslute path to the hmmbuild binary


(optional) A matrix with pairwise alignment scores. For example created by pairwiseAlignment. If the matrix is not provideded by the user it is calculated automatically by the function (time consuming). The sequences are extracted from the alignemnt object.


PSEUDO-ALGORITHM (According to Hannenhalli and Russell (2000)):

  1. (If score matrix is not provided) Run pairwise alignments for all available sequences in the input MSA and save scores to a matrix

  2. (If score matrix is not provided) Calculate a distance matrix based off of the alignment scores one

  3. Perform hierarchical clustering on the distance matrix (UPGMA method)

  4. Get the sequence clusters

  5. Divide the alignment into sub_groups which are the clusters

  6. Run hmmbuild for whole_alignment without sub-group and sub_group

  7. Calculate relative entropy using these two as indicated in the Reference and repeat for each sub_group

  8. Calculate the cumulative relative entropy

hmmbuild program:

This function uses hmmbuild program of HMMER suite for HMM profile generation for MSA.

We recommend downloading and installing HMMER by following the instructions and steps in the HMMER installation website .



A vector of length equal to the length of aligned sequences


Michal Stolarczyk & Alicja Pluciennik


Hannenhalli, S. S. & Russell, R. B. Analysis and prediction of functional sub-types from protein sequence alignments11Edited by J. Thornton. Journal of Molecular Biology 303, 61–76 (2000).

See Also

consensus, cons2seqs_ident, read.alignment


#No example due to external software requirements

[Package BALCONY version 0.2.10 Index]