DNA_NN_SantaLucia_2004.
The TA/AT step was -20.4 (duplicated from the
adjacent AT/TA row) and should be -21.3; the
GG/CC step was -19.0 and should be
-19.9. Both values are shared with
DNA_NN_Allawi_1998, which already carried them correctly,
so the two tables were internally inconsistent. Tm values computed with
DNA_NN_SantaLucia_2004 will change slightly for sequences
containing these steps.Added 19 nearest-neighbor parameter sets derived by melting-temperature optimization (Weber and colleagues, UFMG), bringing the total from 8 to 27. Unlike the existing sets, these are fitted directly at a stated sodium concentration and are intended to replace salt correction rather than be corrected.
DNA_NN_Weber_2015 (1020 mM) and a salt series
DNA_NN_Weber_OW04_69, _119, _220,
_621, _1020.RNA_NN_Weber_VIF_* and
RNA_NN_Weber_FIF_* at 71, 121, 221, 621 and 1021 mM. The
VIF (variable initiation factor) sets gave better cross-validation in
the source study.RNA_DNA_NN_Weber_2019_FT and
_VH (1000 mM) and RNA_DNA_NN_Weber_2019_LS
(100 mM).Values were taken from the parameter files distributed with VarGibbs
5.0 at full precision rather than transcribed from the published tables.
The transcription was validated by confirming that the reference files
shipped alongside them reproduce the existing
DNA_NN_Allawi_1998, DNA_NN_Sugimoto_1996,
RNA_NN_Xia_1998, RNA_NN_Freier_1986 and all 87
rows of DNA_IMM_Peyret_1999 exactly.
Parameter sets now carry a salt_mM attribute when
they were fitted at a specific sodium concentration.
tm_nn() uses it to avoid double-counting the ionic
contribution:
Na matches the concentration the set
was fitted at, salt correction is skipped;Existing parameter sets carry no such attribute, so this is a no-op for all previous usage.
salt_method gains a "none" option to
disable salt correction explicitly. The documentation previously stated
that NA would do this, but match.arg()
rejected it.
tm_nn() and tm_calculate() now report
Salt correction applied (logical) and
Parameter set fitted at [Na+] (mM) in the returned
options, so the automatic skip is visible rather than
silent.
tm_nn() gained a @return section; the
return value was previously undocumented."SantaLucia1998-2" from the documented
salt_method options in tm_nn() and
tm_calculate(). It was listed in the help pages but absent
from the function’s accepted values, so following the documentation
produced an error.?tm_nn
and a “Salt handling” section to ?tm_calculate.@details for tm_calculate() now explains
when each of the three methods is appropriate, including the note that
nearest-neighbor parameters are calibrated on short duplexes, so values
computed over long sequences or fixed-width genomic windows are best
read as a relative measure of local thermodynamic stability rather than
as an absolute experimental Tm.tm_nn() and tm_calculate() now
cross-reference each other via @seealso.