Using principal component analysis to find correlations between loop-related and thermodynamic variables for G-quadruplex-forming sequences.

The application of Principal Component Analysis (PCA) is proposed here as a simple means of revealing correlations between thermodynamic variables corresponding to folding equilibria of intramolecular G-quadruplexes and Watson-Crick duplexes, and the length of loops in the corresponding guanine-rich DNA sequences. To this end, two previously studied data sets were analyzed (Arora and Maiti, J. Phys. Chem. B. 2009 and Kumar and Maiti, Nucleic Acids. Res. 2008). All of the sequences considered shared the common structure 5'- GGG - loop1 - GGG - loop2 - GGG - loop3 - GGG -3'. PCA of these data sets supported a series of correlations between the variables studied. First, the association of loop length with thermodynamic stability and quadruplex structure was corroborated. Secondly, it is proposed that the addition of ethylene glycol produces a stronger stabilization on those sequences showing long loop1 and/or loop3. Thirdly, it is proposed that a low content of adenine in loop1 and/or loop3 will produce an increase in the stability of G-quadruplex and its related Watson-Crick duplex. Copyright 2010 Elsevier Masson SAS. All rights reserved.