[BioC] GCRMA and background
b.otto at uke.uni-hamburg.de
Thu Feb 9 16:14:23 CET 2006
Hi Jim, Mikhail,
JIM> Affy doesn't use a dye like cDNA arrays
I would agree on that...
JIM> No, it doesn't have anything to do with affinities to the dyes
Maybe it should be added that according to Naef and Magnasco the
biotinilation impedes the binding which is in return of interest because
only U and C are biotinylated. Would you agree on that Jim or did I
misunderstand something? Richard, have a look at "Solving the riddle of the
bright mismatches: Labeling and effective binding in oligonucleotide arrays"
by Naef and Magnasco, that is the paper I think is discussing the issue.
In addition to the three versus two hydrogen bond aspect, again according to
the paper mentioned above, the affinity has to do something with the size of
the dNTPs. Note that we have pyrimidines (U,C) which are smaller in size
than the purines (A,G). When you have a mismatch somewhere in the binding
sequence ( some nonperfect match) the affinity of the binding is highly
depending on how the mismatch looks like: if two small bases come to stand
against each other, they just dangle and you only loose the binding energy
of the hydrogen bonds. However if some purines face each other you get a
steric problem. Based on these findings, on page 3 of the paper you get a
potential explanation why we sometimes observe higher mismatch signals than
perfect match signals. However I'm not quite sure how these aspects are
directly incoparated into GCRMA. I just remember that Zhijin referred in
their paper to the Naef+Magnasco one.
More information about the Bioconductor