Correlated variation
13 Dec 2011
In predicting functional sites in proteins, conservation of amino acids in sequence alignments is a well-known indicator of functional properties. In addition, correlated variation among amino acid positions tells something about which residues are located close to each other in the 3D structure can be applied for the prediction of intra- or intermolecular contacts.
Current methods are limited to the analysis of pairwise correlations, but this obscures the difference between direct and indirect correlations. Observed correlation therefore thus not necessarily indicate that residues are located close to each other.
In a paper in BMC Bioinformatics, Sreekumar et al. present a new method for Regularized Multinomial Regression to obtain Correlated Mutations (RMRCM) from protein multiple sequence alignments. Using simulated and biological datasets, good performance of the method was shown.
RMRCM is available in R-code via www.ab.wur.nl/rmrcm
J Sreekumar, CJF ter Braak, RCHJ van Ham, ADJ van Dijk
Correlated mutations via regularized multinomial regression
BMC Bioinformatics 2011, 12:444
Highlights
The winner of the NBIC Young Investigator Award 2012 is Dr. Christian Gilissen, UMC St Radboud, Nijmegen....
e-Infrastructure colloquium
Many software projects in the life sciences are written as "open source". But do we actually kn...
This course covers the recent advances in discovery informatics, with a focus on the application of e-sci...
NBIC, SIB, EBI and BTN (Bioinformatics Training Network) organise a one-day workshop related to bioinform...