[meltsim] Re: Question regarding MeltSim

J.W. Bizzaro jeff at bioinformatics.org
Tue Mar 21 17:56:14 EST 2006


Hi Christine,

MeltSim's empirical parameters include stability parameters for all 16 
nearest-neighbor pairs (instead of the 10 unique pairs), which are all assumed 
to be on the same strand.  From the source code:

#define GG_STAB salt_log10 * 14.18 + 391.65
#define CC_STAB salt_log10 * 14.18 + 391.65
#define GC_STAB salt_log10 * 13.20 + 397.77
#define CG_STAB salt_log10 * 13.20 + 397.70
#define AC_STAB salt_log10 * 16.21 + 381.12
#define GT_STAB salt_log10 * 16.21 + 381.12
#define CA_STAB salt_log10 * 17.10 + 376.34
#define TG_STAB salt_log10 * 17.10 + 376.34
#define AG_STAB salt_log10 * 16.87 + 377.59
#define CT_STAB salt_log10 * 16.87 + 377.59
#define GA_STAB salt_log10 * 17.76 + 372.65
#define TC_STAB salt_log10 * 17.76 + 372.65
#define AT_STAB salt_log10 * 21.00 + 355.01
#define TA_STAB salt_log10 * 20.11 + 359.88
#define AA_STAB salt_log10 * 19.78 + 362.24
#define TT_STAB salt_log10 * 19.78 + 362.24

But, since an "AT" NN pair, for example, has a different stability than the 
reverse "TA" NN pair, you'll get different amplitudes (albeit somewhat similar 
features) between derivative plots, depending on the direction in which the 
sequence is read.

Notice, though, that the stability parameters are the same between NN pairs 
that are the reverse *complement* of each other, meaning it doesn't matter 
which strand you use.

Cheers,
Jeff

Christine Kingsburry wrote:
> Hi,
> 
> My name is Christine Kingsburry and I'm working in Gary Slater's group at
> the University of Ottawa, Canada.  We recently got interested in DNA
> melting and in your program MeltSim.  We noticed that there is discrepancy
> between the curves obtained for the same DNA sequence, but in the inverse
> form.  We were just wondering if you knew the reason why.  Is it a 3' 5'
> effect?
> 
> Thank you
> Christine

-- 
J.W. Bizzaro
Bioinformatics Organization, Inc. (Bioinformatics.Org)
E-mail: jeff at bioinformatics.org
Phone:  +1 508 890 8600
--


More information about the meltsim mailing list