dimer step parameters calculation for a DNA with flipped bases

[cosials]

Dear Dr. Xian-Jun Lu,


I'm using 3DNA to calculate the dimer step parameters from a protein-DNA structure. The DNA has a slightly bent conformation with two bases at the central region that are completelly flipped in a hextrahelical conformation.
While running your software, the program do not calculate the dimer step parameters where the flipped bases are allocated. See attached the table with a gap in the middle section. Moreover, you can see the plot with the roll and twist parameters. Below the plot, there is the DNA sequence. Asterisks indicated the flipped bases.

Have you ever seen something like that? Is it because the presence of a flipped base makes impossible to calculate the dimer step parameters? Should I omit thie data points at the central region?

Thanks a lot,

Anna

[xiangjun]

Have you ever seen something like that? Is it because the presence of a flipped base makes impossible to calculate the dimer step parameters? Should I omit thie data points at the central region?

Yes, such distortions are common in DNA-protein complexes (e.g. complexes of DNA methyltransferase with DNA). The flipped base does not form a base-pair anymore, as can be seen clearly using Jmol/PyMOL, and thus not reported by the 3DNA find_pair program.

You could force the 3DNA analyze program to calculate step parameters involving flipped bases. See FAQ entry How to fix missing (superfluous) base pairs identified by find_pair?, and the 3DNA 2008 Nature Protocols paper. However, such parameters do not make 'intuitive' sense, even though they can be used to rigorously rebuild the relative base geometry of the original structure.

You may simply omit the point in the central region, with some note in the text

HTH,

Xiang-Jun